Leslie Carr and
Multimedia Research Group,
University of Southampton.
Authors' note. Web publishing poses a dilemma for authors of academic papers: speed of publication against the requirement for refereeing or validation. This is an unrefereed paper but it is CURRENT (at least, to within a few months of the latest date at the end of this note). This is the essence of Web publishing, the speed of delivery and the ability to access related materials. As long as the paper is current there can be confidence that the links it contains - there are over 100 in this paper, and every journal covered in the survey is linked - will be effective, so click away.
The paper is in two parts: a text-based commentary and a series of tables, some extensive. The two can be used fairly independently. Table 6, for example, contains all the journal links. It is recommended that the paper is read online, so that links can be followed, using a tables-aware browser such as Netscape 1.1 or later. If you wish to download the paper prior to reading, note that the longer tables 2, 3, 4 and 6 are held in separate files. Each one has to be viewed and downloaded separately.
If you are averse to unrefereed papers, we are considering placing an abridged version in a refereed journal, but it is unlikely to have links. There is an alternative. We have posted relevant and constructive comments on the paper in edited correspondence files, a sort of open peer commentary. Please email views that might assist authors or readers to firstname.lastname@example.org.
FAIR USE: We reserve the right to distribute this electronic document. It is publicly posted to the World Wide Web, and anyone is free to establish a link from it to a subsidiary server, but not to copy it for public posting on a remote server, since that could lead to an undesirable proliferation of obsolete versions.
15 January 1996 (Updated 14 February) This paper has been downloaded times.
Counting electronic journals is not easy, or so it appears. The prevailing perception is that the number of scholarly electronic journals is growing rapidly, and while for science titles there was some truth in it in 1995 if not before, this still creates a misleading impression. Judged in the wake of the phenomenal growth of the World Wide Web, given the demonstrated power and growing influence of the preeminent scholarly electronic publication, the Los Alamos physics e-print archive, and in relation to the number of paper-based scholarly journals, the number of scholarly science journals available online towards the end of 1995 was tiny.
All this is about to change. Already a number of UK-based academic publishers have announced their intention to make all their journals, hundreds of titles overall, available online during 1996. The principal enabler for some of these publishers is a pilot scheme set up by the Higher Education Funding Councils for the UK (Bekhradnia 1995) in which the publishers are guaranteed a portion of their existing subscription income directly from the councils in return for which the publishers have agreed to deliver both paper and electronic editions to subscribing UK libraries, which incidentally will pay less for these subscriptions than at present. Coincident with the resolution of the subscription issue is the publishing industry's apparent convergence on Adobe Acrobat (Brailsford and Harrison 1994), one of a number of portable document formats that allow page images to be transmitted and received electronically without loss of content or formatting data, without noticeable image degradation and without requiring expensive software viewers - in fact the viewers are free. In other words, publishers are finding ways to increase their income sufficiently to cover the marginal additional cost of generating an electronically-deliverable format from the original journal pages at the same time as the underlying technology required to implement the process has begun to reach maturity both technically and in terms of market acceptance.
As a first step this approach is justifiable from the viewpoints of publishers and, initially, of users, who will benefit from the flexibility of electronic access, the ease of printing selected pages and who will find comforting familiarity in the appearance of documents they receive. It is not clear, however, that either the technology adopted in this process or, more fundamentally, the underlying publishing philosophy are sufficiently extensible to embrace the subsequent demand for all-electronic journals. By their nature such journals must, to exploit the medium to the full, eventually diverge from their paper counterparts and require a new and radical conception of the journal as an interactive, hypertext-linked, multimedia product, with the ramifications that will become apparent in the paper.
Before being submerged by page-based electronic journals, which the paper refers to as electronic editions, this survey examines the publishing features of online journals which have evolved through a more random process and which may hold clues to the eventual, longer-term shape of the electronic journal. What are the most innovative and useful features in the best online journals, how widely are these features being adopted, and what new directions are apparent from these real implementations?
Scope of the survey
As a means to produce practical and meaningful results, the survey concentrates on English-language online journals in the fields of science, technology and medicine (STM). The term `STM' is used as a broad market definition by many publishers of paper scholarly journals. Paper journals in these fields tend to be among the most expensive of all scholarly journals and therefore would seem to have most to gain, from the purchaser's point-of-view, through any cost reductions that electronic publishing could achieve (Duranceau et al. 1995). STM journals contain specialised terminology, they frequently include detailed mathematics and often have complex artwork and tabular data, and so are usually regarded as the most difficult journals to produce, in any form. It might follow that if it is possible to produce such journals in an electronic format, then the die will have been cast for others to copy.
The only other requirement for journals to be included in the survey was that there is evidence that full papers from the journal are available online - but not exclusively online - at the time of the survey, which was carried out during September and October 1995. Restricted access, typically through subscription requirements, meant that not all journals could be viewed. At the present stage of development it is invariably only commercial publishers that restrict access to online journals. Conforming online journals that had been announced but which were not available at the time are not covered in the survey.
Paper scholarly journals typically have the defining features that they report full and novel research results, are supported by literature references and are subjected to rigorous peer review. It was not a requirement of this survey that online journals have these features other than that published papers are subject to some form of refereeing, although for this survey this is to differentiate research-level publications from news-oriented journals rather than implying any measure of the quality of the review process of any particular journal. Refereeing of some sort, it is widely regarded, will be indispensable if online journals are to claim large numbers of users. It is not clear that novel results should be a prerequisite of papers published in online journals, in the sense of the strict criteria on originality that currently apply to the submission procedures of many paper journals. One major STM publisher, the Association of Computing Machinery (ACM), for example, with reference to the growing practice of circulating unrefereed preprints, notes in its electronic publishing plan that `this obviously poses a challenge to the policy of not considering previously published works' (Denning and Rous 1995). Methods of referencing, especially through hypertext linking, will be enhanced dramatically with online journals (Open Journal Project 1995).
Electronic journals can be differentiated according to delivery formats. Woodward and McKnight (1995) identified three formats for electronic publication: online, CD-ROM and networked. This survey only includes what the authors described as online and networked journals, although here referring to them all as `online', that is, available from the Internet, invariably through the World Wide Web. For scholarly journals, only online delivery enables the radical journal referred to above. The survey does not cover electronic journals available only in CD-ROM format, some of which were mentioned by Woodward and McKnight.
Issues considered are novel electronic features, publishing history, the style of publications in different subject fields, formats and additional delivery methods, the volume of material available electronically, which is especially telling for many new online journals, the impact of new non-commercial producers as well as the response of well-known commercial publishers, charging practices and future plans. Full reference details of all the journals included in this survey, including URLs for access via the Web, are given in the Appendix.
It would be wrong to claim that the survey includes every journal available online within the category defined. Some publishers, notably the multinational Elsevier (1992), have been making page images from journals available online for some time and would claim in this sense to be an electronic publisher. Approaches such as this are often highly restrictive with non-subscribing individuals denied access even to samples or experimental issues, so judgments are difficult. A contrasting approach is that of the BioMedNet information service (1995) through which a large number of biology journals are available in screen-oriented formats with innovative electronic features. The service has been available free on a trial basis but from 1996 will require a subscription both to the service and to the individual journals. To avoid distorting the overall survey towards one publisher's preferences the full list of journals available from such sources are not included in the figures.
Nor is the Los Alamos physics e-print archive part of the analysis presented here. It could be considered a single publication or a collection of journals; in fact it is not considered to be a journal, in the conventional sense, by its originator, Paul Ginsparg, who prefers to distinguish the archive from its paper counterparts. Inclusion of the archive would, as above, dominate the results of the survey, possibly masking other features. The significance of the archive is much greater than any single journal, and its impact and implications are considered in a related paper (Hitchcock 1995b).
The broad picture
During the period of the survey 115 STM online journals were accessed through the Web if the BioMedNet journals are included, 83 if they are not. Other than the broad comparison made above with the number of paper journals, no particular significance should be attached to these figures since they are merely a function of the instant of the survey and depend on how the journals are counted. It is notable, however, that at least 44 of the 83 journals first appeared in 1995. Prior to that there were a mere 13 electronic-only journals, the only journal type that can be reliably dated from Web pages, although of course this does not include the electronic editions that were online at the time. Another perspective on the growth in the number of electronic journals over the last year or so, and on the changing culture of such journals, is afforded by comparison with a similar survey by Roes (1994).
Further scrutiny of the 83 titles shows that the majority, 48 journals, are electronic editions (Table 1). A significant number of journals, 36, require, or will require in the near future, a paid subscription to view. This latter figure especially is higher than might be expected, to judge from similar surveys. Harter (1995), in a study for the Online Computer Library Center (OCLC) in the USA of 125 electronic journals, which included 76 peer-reviewed journals, has indicated in preliminary findings that almost 90 per cent of the electronic versions of these journals are free. Harter says that his survey includes almost equal numbers of journals in the sciences, social sciences, humanities and professions. The higher percentage of fee-based journals found in this survey reflects the emphasis on technical, specialised journals that have a need to recover some of the higher costs than might be found more generally among electronic journals or electronic magazines.
STM online journals come from three main sources: commercial publishers, mostly well established and familiar names; learned societies which are themselves invariably highly commercial, if non-profit, organisations; and other research institutions, typically, but not exclusively, universities. The latter might suggest university presses, but this is not the case, probably because the presses generally have a limited interest in the sciences. Instead, within this sector online journals are often initiated by individuals, departments or special-interest groups, sometimes across faculties or between different universities.
These three sources are fairly evenly represented as producers of online science journals, with the commercial publishers producing, by a small margin, most titles. Two-thirds of paid-for titles in the survey involve commercial publishers, although it is not true to say that all titles from this sector are fee-based. More predictably, all but a single title from the university-based sector, that is those groups that produce journals with no intervention from commercial agents, offer free access for readers. The figures for all three sectors are shown in Table 1.
Clearly this picture will not survive far into 1996. Once the HEFC-supported journals come online, paid-for commercial titles will dominate. One advantage for UK-based readers of these journals is that papers are likely to be freely accessible from any computer that has a network connection within a subscribing establishment. In other words, despite the paid-for status these journals will appear to be free to most users in the UK. This contrasts markedly with some other commercial arrangements for subscribed online journals which restrict access to defined terminals, limit the number of simultaneous users and prescribe the permitted volume of copying or printing, often with potentially horrific and unpredictable costs to the subscribing library or institution.
Novel electronic features
A range of innovations and features become possible with online journals that are either impossible or difficult to achieve with paper journals or even with electronic editions. Since in the first instance those that new electronic-only scholarly journals need to attract are authors rather than readers, the availability and accessibility of these features are widely believed to be more important, currently, in motivating the transition from paper to electronic journals than the cost issues, which typically have little impact on the user as author.
What features? A random glance at an online journal in this survey is likely to reveal few innovations. One feature is that electronic journals can be searchable, and some are. Papers can include hypertext links to other resources. Many biology journals make particular use of this facility to link to relevant databases, but few other journals offer links to resources beyond the journal. Colour graphics and photographs are not as costly to include in electronic journals as they are to print on paper, but they are rarely used, and video and audio, unsurprisingly, are to date even rarer. Electronic editions can present materials in advance of paper publication, and many do, although unedited preprints are not common.
Currently one of the best examples of the use of these features in an online journal is Astrophysical Journal Letters, launched in July 1995. This presents the letters section in advance of publication of the full paper journal. The growing archive is searchable, and in html screen form papers have hypertext links from thumbnail figures and photographs to full images, threaded links to create themes and support subject categories, and external links to a database of abstracts. Papers are also available in pdf form for local printing.
Readers of online journals can be alerted to new papers or issues via electronic mail, discussion lists or newsgroups. Clearly this is not exclusive to online publications, but such an information service presupposes that the recipient is online and therefore seems to carry more weight if the publication is also online. Importantly, these services can be extended to embrace higher degrees of interactivity, open peer commentary or 'scholarly skywriting' for example (Harnad 1990), which is greatly more effective for online publications. A handful of journals in the survey have one or more such services.
That these features are sparsely distributed among the current generation of online journals is evident from Table 2, which appears in two parts for electronic-only journals and electronic editions since this distinction has a bearing on the features that are possible. Table 2 also indicates for each journal the amount of material that is available online. In many cases a journal has to become established and active before the electronic features evolve.
In this latter respect there is an obvious difference between the two journal categories. For electronic editions it is for the publisher to decide how many papers from the current issue and how much of the back archive to make available online, while for many electronic-only journals the papers published in a whole year would not fill a single issue of a paper journal. For these electronic-only journals it is a classic Catch-22 case: no features, no papers; no papers, no features. This emphasises the need, for online journals as much as for paper journals, for fundamental editorial judgments, for differentiation and the need to identify a competitive edge regardless of technology.
There have been some electronic-only successes. Some of the early online journals in mathematics have become established, as in computer science has the Journal of Artificial Intelligence Research from 1993, all with a steady if moderate flow of papers, but it is doubtful that any has had a noticeable effect so far on the paper-based competition. Psycoloquy, in 1990 one of the first online journals and perhaps one of the best known examples of the genre outside its specialist field, has evolved to encompass new features: from plain text, which is still supported and archived, to the more flexible html version with hypertext threads between papers. The journal has pioneered open peer commentary online, and readers are notified of new papers, all with no restrictions on access. Strictly, nothing is made available on paper.
Current journals have hardly begun to implement the novel features that online delivery makes possible. Online journals will develop and improve, but a difficult decision has to be made by publishers and editors of prospective new journals if they are to realise these improvements: publish an edition on paper, or remove paper from the process entirely. Paper journals are the only commercially proven medium, and established paper journals continue to attract the majority of papers and the best papers (Harnad 1995). Paper journals are a constraint on the features of `parallel' online journals, however. A publication predicated on the paper journal model cannot have real interactivity at online speeds and multimedia content, for example, without diverging from the paper production process and content. Are these features ultimately to be denied? Clearly not, but there is as yet no obvious path from a paper or `parallel' publication to an electronic-only publication, and as long as this remains so publishers and editors will increasingly be faced with this dilemma.
Presentation formats: a timeline for online journals
From the first online journals at the beginning of the decade the format of an online journal has typically been a product of the subject of the journal, the time of its first appearance and, once again, whether it is an electronic-only journal or an electronic edition. Table 3 is organised to demonstrate this.
Mathematics has generated the largest number of online journals of any field. The first maths journals appeared online in 1993, and the majority are electronic-only. All but two of these journals are available in Postscript, a format designed for page printing but which can be produced from copy supplied in the mathematician's formatting language TeX. So far only two, more recent, online maths journals provide the additional option of distilling the Postscript copy into Adobe's related but more flexible pdf (Smith 1995). Most of these journals also provide the option to view in TeX or dvi formats.
Physics journals make similar demands on mathematical formatting capabilities, but owing to the dominance of the Los Alamos e-print archive there is just one electronic-only journal and a number of electronic editions, almost all of a recent vintage. This, and the fact that most electronic editions in this field are the half-hearted attempts of learned society publishers playing catch-up with the e-print archive by extracting letters and short communications rather than full papers from the journals, has resulted in a greater variety of formats, including html and pdf as well as some examples of Postscript, TeX and dvi.
Medicine and biology are both well represented with online journals. Apart from The Online Journal of Current Clinical Trials from 1992, the earliest journals in these areas appear to have begun in 1994, and combined with the lesser reliance on presenting mathematical formulas but the strong dependence on graphics, html is almost uniformly the format of choice, with some publishers of electronic editions preferring pdf. There is one major difference between the two fields. While a number of major biology titles are available online, none of the most prestigious medical journals are available in this form. In biology the number of electronic editions has thus reduced the scope for new electronic-only journals, but in the medical area electronic-only journals are in the majority.
It would be hard to claim significant impact of online journals in any other field. At the more technical end of the spectrum Postscript dominates: the Journal of Universal Computer Science has temporarily withdrawn an option for html because of problems with presenting mathematics, but is still available as Postscript. The few online journals in the more general areas of information technology and education invariably use html.
Some other formats of limited remaining relevance require comment. Plain text, or ascii, was the most universal format available to the earliest online journals, and it still has an archival role for these journals. Today it has its advocates as an optional format that can be viewed on the most basic systems and which may be important if journals are to be accessed in developing countries, but it does not have the required sophistication to represent most scientific materials.
Some argue that as a stable and standardised format ascii should be used to archive many more online journals. For the same reason, a case is made for storing archived paper journals as scanned bitmap images, with the anticipation at least of being able to view the required pages in, say, 50 years time. For practical purposes, especially for electronic-only journals and even for electronic editions, bitmaps serve no purpose other than archival backup. In this survey only the journal Bioscene is presented in this form. The journal has the largest online archive of all those in the survey, stretching back 21 years, but the image quality is poor, which the publisher appears to acknowledge. Elsevier's TULIP project (1992) is also a major user of this format.
Rich text format (rtf) is used by a single journal, the Annals of Saudi Medicine, although this journal is also available as html. As a format that has a substantial user base through its connection with Microsoft Word and which is widely used for document transfer, rtf nevertheless fails to deliver the range of presentation features necessary for typical STM journals and is not a sufficiently standardised cross-platform format for publishing applications.
The remaining format to appear in Table 3 is Guidon. Strictly it is not a format but a graphical user interface (GUI) developed by OCLC for the delivery of journals that use its Electronic Journal Online service, but it is included here as it controls the display of the journals viewed through it. Guidon is a Windows-only GUI which supports figures, tables, equations, full-text searching and local printing. Five journals in this survey can be viewed through Guidon, although some of these can use other interfaces such as Netscape Navigator.
It may seem odd, heretical even, not to mention sgml. Despite its undoubted long-term importance to electronic publishing as a storage format, it is not currently a presentation format for any journals covered here. Its influence, however, is pervasive. HTML is a specialised sgml document type definition (DTD), a particular instance of sgml. The relative simplicity of the html instance of sgml is both its strength and its weakness, but while html has been extraordinarily successful at establishing itself as the language of the Web, the debate on its further development has to consider how to extend its capabilities for professional publishing applications without losing the simplicity that has made html so popular. If the resulting html fails to represent, say, mathematical or chemical symbols adequately, more information-rich sgml DTDs will need to be developed.
Stimulating electronic journals
With a view to longer-term remodelling of the journal as an electronic-only product, on the basis of the survey it may be worth speculating on the motivations for developing such journals. Given the stronger representation of some fields with electronic-only journals, processes must be at work to enhance the desirability and usefulness of this type of publication in these fields.
Physics had a preprint culture, in which papers are circulated in advance of refereed journal publication, some time before electronic distribution became viable (Ginsparg 1994), but once it did it was a natural medium for physicists to exploit for preprints. Maths similarly has developed preprint services, and a long history of comprehensive review journals in this field suggests a concern with an organizational framework that improves the awareness and accessibility of literature within the field. As has already been mentioned, the necessary formatting tools such as TeX were available to enable the complex features typically found in papers in these subjects to be viewed effectively and reliably.
If the use of TeX were a prerequisite for authors of papers to be published online then it is unlikely that the culture of electronic publishing would have spread much more widely than these two fields. As more amenable environments for the creation and presentation of electronic papers have emerged, however, more disciplines began to exploit them. Biologists required access to graphical, often three-dimensional, representations of molecular structures. They recognized the potential to compile databases of these computer-generated images and to make them freely available online. These databases in turn have become a resource that online journals can exploit by linking directly between journal papers and items in the databases. It is this `interconnectedness' that is clearly a motivating factor.
The same factor applies in medicine, although here the advent of electronic publishing is embraced more reluctantly, as a number of high profile newspaper articles and editorials testify (New England Journal of Medicine 1995). Medical researchers are unwilling to trust critical results to a new medium without the confidence that long-established verification procedures can be reproduced through this medium and will not be prone to abuse or malpractice. Despite this, the field has its advocates, practitioners who have championed electronic publishing in prestigious sources (LaPorte et al. 1995), and has gone on to produce a reasonable number of journals. It appears to be one of the most active fields for new developments.
It may be surprising that computer science is poorly represented by electronic journals, but the reasons may now be clearer. None of the motivations identified above apply strongly in this field. There are a number of online ftp archives of computer science reports, for example the Universal Computer Science Technical Reports Index at the University of Indiana (VanHeyningen 1995), but there is no organised preprint culture and wide distribution of papers in this form appears to be less important than in physics. Computer science has a conference culture which appears to be migrating online, although in an ad hoc rather than any focussed way. Computer scientists can handle the formatting tools perfectly well if required, but appear to have little need or desire to do so. The computer science world is certainly physically connected, but the organised information `interconnectedness' displayed elsewhere has not yet developed to the same degree since there are no focal points such as major databases and few established connection structures that can be mimicked and developed online. Where are the eminent and influential computer scientists who are prepared to champion new electronic publications in the face of the current dilemmas, doubts and divisiveness?
In the longer term electronic publishing may be an agent to change the boundaries between academic disciplines. More immediately online journals will be subject based, and in any given field the motivations for new online journals will depend on a shared vision of what is possible and the ability to build on established cultures within that field. Where these motivations are unclear, electronic-only journals if not electronic editions will be slow to materialise.
Funding electronic journals
There will always be a cost attached to the production of online journals, even though a glance at Table 4 might suggest otherwise. How that cost is absorbed is crucial to the future of online journals. Should the cost be borne by the reader or by a subscriber on the reader's behalf? Should it be contributed by the author? The survey reveals that a number of ways are being used to pay for online journals, with the subscription model prevailing strongly for electronic editions, but also points to possible instability for some electronic-only journals.
On the issue of funding, a clear distinction again arises between electronic-only journals and electronic editions. Table 4 is structured to highlight this. Within this framework journals are grouped by types of producer, and then according to whether they are, and will continue to be, free to the reader.
There is little chance that publishers will forsake the long-established practice of generating income from subscriptions. It is fair to assume that most electronic editions will require a subscription to view, even if that subscription is in essence to purchase the paper copy. Subscribers will increasingly be offered more options to purchase one or more versions on paper, CD-ROM or online, as is currently available for example for Applied Physics Letters Online (1996). There will be those who are sceptical about paying more for another copy of the same material; if for an additional version the extra cost is in the region of 15 per cent, as it is in some cases with APLO, this may be justifiable, but not perhaps up to 40 per cent as is also evident. One publisher at least, although its journals are not covered in this survey, charges less to subscribe to an online copy than to the paper journal. Online journals from the Muse project of Johns Hopkins University Press (1995) are available for 10 per cent less than on paper.
A more radical approach is `pay-per-view', where readers might be requested to bill an account each time they elect to view, or download, a paper from a given journal or publisher. There is no evidence yet that this will be widespread. The structure of viewing arrangements on BioMedNet suggests that it is considering this option. During the trial period users were asked, for each paper selected via a button link, to authorise a payment of $0.00. Cambridge University Press, publisher of Protein Science Online, also appears to be considering this approach. The ACM's electronic publishing plan (Denning and Rous 1995) makes clear reference to pay-per-view, although no journals are yet available to assess.
While some journal users express a willingness to be charged on a pay-per-view basis, this is typically at penny rates and may not encourage publishers into this unpredictable way of financing journals. There must also be some concern that a pay-per-view scheme that charges significantly more, say in the region of current copyright clearance fees of a few dollars, will adversely affect one of the principal benefits of online electronic publishing, that of linking between papers, journals and other resources. Links will be less appealing if each one has a price tag attached to it. Speculatively, one might see the growth of link publishers, using new approaches such as the Distributed Link Service (Carr et al. 1995), who will buy the rights to access linked materials, while charging users a single fee or subscription for a continually updated link service. The commercial objective of a link publisher must be to enable beneficial use of links on such a large scale that the volume of use creates a cost-effective environment for link development.
It could be argued that to realise the full potential of scholarly online journals, especially in terms of interactivity or linking, they must be free to the reader. At the very least, the costs or barriers to the reader's use of journals must be reduced. One way of doing this may be author page charges (Harnad 1994). This has long been an accepted method of funding paper journals in some fields, although the practice may be less popular than it was. Just two journals in the survey impose author page charges to fund publication, the electronic-only journal InterStat and the learned society electronic edition Florida Entomologist (FE). Another learned society, the Optical Society of America, is considering page charges for Engineering and Laboratory Notes. These societies apart, there is little evident motivation for page charges, and it appears this would be unlikely to change without some fundamental change in the funding provisions for research, or much lower page charges than have been common or are being applied now to FE.
Some electronic editions will continue to offer free access, but this will either be for an experimental or promotional period, to selected samples or parts of the full journals.
Electronic-only journals, especially where they are produced outside a commercial framework, would appear to have a better chance of providing a full service free of charge, but this is misleading. Some good journals are being produced at very low cost, but this could not scale to support the publication of all scholarly papers. It is not realistic to expect the necessary people to work, long term, on these journals at no cost and achieve the required standard or quality. There will be a cost, even if it is substantially lower than the cost of producing paper journals, and ways must be found to cover it (Hitchcock 1995a).
Apart from page charges, Table 4 shows some electronic-only journals supported by research or educational grants, and some by appropriate associations or societies, but by its nature this type of funding is often short-term, and the gap still has to be filled. A popular idea is for the non-commercial producer to liaise with a commercial publisher to publish an annual compilation of online papers, on CD-ROM or even as a bound paper volume. Reverting to paper seems to be a backward step for electronic-only journals, and the market for CD-ROM collections of `old' papers is unproven: there is obvious archival value, but not at the price level of current journal subscriptions. One publisher, Priory Lodge Education, hopes to generate sufficient advertising or sponsorship to support its online medical journals. This may be feasible in the medical field, but it is unlikely that online journals could generate advertising in fields where none existed before.
The danger for electronic-only journals is the proportion of current journals, just under half, that have no officially stated way of raising any funds. This could become destabilising for this type of journal, and viable solutions must be arranged if new electronic-only journals are not to be discouraged, particularly in view of the healthier financial state of subscription-based journals. For many, though, this type of funding is not the objective, and subscription journals equally face the severe problem of the worldwide library funding crisis (Okerson 1992).
Experimenting with the structure of journals
Another way in which online journals could diverge from paper journals is in their structure and frequency of publication. Papers can appear online as fast as they can be editorially approved, or instantly if preprints are accepted, rather than in periodic bundles, and the number of papers made available is constrained, if at all, by quality criteria rather than page limits. There is the opportunity to develop more broadly-based journals, reversing the process of fragmentation that has in recent times dominated new paper journal publishing.
Two journals are worth noting in this respect. The Journal of Universal Computer Science takes a highly broad-brush approach to its subject. Whether this is useful remains to be seen, but it seems likely that more targetted strategies will be necessary. A more promising approach involves what the editors call `multiple virtual journals'. InterJournal is a number of interlocking journals on different subjects to which papers can be submitted simultaneously. The interrelated subject areas - currently complex systems, polymers and complex fluids, and genetics - are all part of the same global hierarchy and all manuscript information is stored in a common database (Redi and Bar-Yam 1995). Given the small number of submitted papers, the journal has yet to demonstrate that the idea has wide support.
Online journal awareness services
With the expanding number of online journals and readers new to online access, how can resources, especially new journals, be discovered reliably? Are online journals easy to find? Yes they are providing the searcher is aware a journal exists, especially if they are available over the Web. If not, and this particularly affects new electronic-only journals rather than electronic editions, the range and growing sophistication of the now largely commercialized Web search services may provide some answers, but these searches are dependent on the user's query and keyword selection. An alternative is to subscribe to, typically electronic mail-based, alerting services, but this option would involve a high information overhead for a user with specialised interests.
Perhaps the best option is to check one of the numerous journal lists that are available on the Web, and these provided links to many of the journals included in the survey. Although a number of lists and alerting services were monitored for relevant journals before and during the period of the survey, none was found to include all the journals discovered, with the best `hit rate' above 60 per cent and the lowest below 20 per cent, even though for most lists the scope is broader than that adopted here. This reflects the methods of information collection and the selection criteria used by each list, many of which are run by individuals within a library service, rather than any inherent or long-term limitation of such services, but there are implications for marketing and launching online journals.
Not surprisingly, because one of its category listings is `electronic journals, STM (peer reviewed)', the WWW virtual library service produced by E-doc includes by a long way the highest number of titles reaching the criteria of this survey. Another major directory, the Association of Research Libraries' Directory of Electronic Journals, Newsletters, and Academic Discussion Lists, derived from the NewJour mailing list which notifies all new electronic journals, strangely provided fewer than half of the titles in the survey. A more comprehensive database appears to be in preparation. Table 5 lists the journal directories visited for the survey, in order of hit rate, and indicates the general likelihood of finding STM journal titles from these sources.
Until a list of journals can approach a 100 per cent hit rate all the time, a common misunderstanding about `marketing' new journals on the Web must be addressed. Marketing has two purposes: to raise awareness of a product, and to promote sales. On the first, it has been suggested that awareness of a product can be generated almost automatically by virtue of its presence on the Web. The tools that have emerged to support the Internet - notification services such as discussion groups or listservs through email, or newsgroups; Internet search tools such as Archie, Veronica, Wide Area Information System (WAIS); or the Web search engines such as Lycos, WebCrawler or Alta Vista, and features such as hypertext links - taken together surpass by some margin conventional methods for raising awareness of products. Even on the Internet, however, it is seen above that no single information service for journals is complete, currently. More comprehensive services will develop, but they will depend on editors and publishers using and informing these services rather than relying on automatic Web `robots'.
Some major publishers, those that have resolved what they expect to be an acceptable way of charging for electronic editions of existing paper journals, are about to move surprisingly quickly to make these journals available online. Innovation is not confined to the major players, however, and the survey has found some of the state-of-the-art online journals being produced by smaller, independent groups.
Some electronic-only journals are beginning to demonstrate and exploit the potential of the new medium. The next stage should see real examples of multimedia enhancements involving sound, video and simulations, particularly in the fields of biology and medicine. The SuperJournal project (1995), which involves over 20 publishers in the UK, is addressing the integration of multimedia content within scholarly journals. Enhanced hypermedia links will proliferate through the use of `open' link services that separate link and document data, removing the dependence for link creation on the original author, an important feature for widely available and archival journals (Open Journal Project 1995).
Ironically, the third key feature of online journals, the 'scholarly skywriting' type of interactivity, is already supported by the technical infrastructure but may have to await new publishing strategies providing enhanced accessibility before it can become widely practised within scholarly journals.
Formatting of technical material, especially maths and tables, in online journals should improve with the appearance of the specification for html 3.0 (HTML working group of the Internet Engineering Task Force, 1995), while more sophisticated screen-based page design, layout and multimedia integration and delivery will be supported by interfaces such as Netscape 2.0 (Netscape Communications 1995) and Microsoft Internet Studio (Microsoft Corp. 1995).
Then there is Sun's Java interface, which has the potential to add defining features to online publications. Essentially a 'computer within a computer', Java may become a universal operating system (Singleton 1996), enhancing usability by hiding incompatibilities between different computer platforms. Java will bring new features to online publications, enabling 1000s of tiny programs or 'applets' to be shared securely over the Web and adding, for example, animations and spreadsheets that analyse incoming data to otherwise static pages.
After the false dawn in the mid-1980s of desk-top `publishing', which really presaged desk-top production, the combination of these interfaces with the online distribution capability of services such as the Web heralds a genuine capability for cost-effective small-scale publishing. The financial instability and muted impact of small-scale, independent electronic-only journals in this survey, however, suggests that a wider network of support services and professional production, if not the support of established commercial publishers, will continue to be necessary to maintain scholarly publishing at its present scale.
The immediate future of online journals is set to be dominated by electronic editions based on established paper journals and retaining the appearance of familiar paper layouts through Adobe Acrobat. The innovative features made possible by online publishing may therefore be obscured for a time. Their emergence on a large scale will depend on whether electronic editions can evolve to become electronic-only journals, or on the commitment of the scholarly community to demand these features through the development of new journals, producers and publishing structures.
These details are shown in Table 6.
Three publishers, Academic Press, Blackwell and Institute of Physics Publishing, are participating the three-year HEFC national site licence pilot scheme. More details of their plans to publish extensive lists of electronic editions can be found at the following URLs:
Table 1. Summary of journals data
Table 5. Source directories for online journals
|WWW virtual library, electronic journals, STM (peer reviewed) http://www.edoc.com/jrl-bin/wilma/spr||Entries are added and maintained through Web Information-List Maintenance Agent|
|ARL Directory of Electronic Publications, 5th edition gopher://arl.cni.org:70/00/scomm/edir/edir95/jz http://www.gold.ac.uk/history/hyperjournal/arl.htm (hypertext version)||An abridged version of the more detailed database begun at the Association of Research Libraries|
|University of Houston Libraries, scholarly journals distributed via the WWW http://info.lib.uh.edu/wj/webjour.html||Points to 'more comprehensive' directories|
|CICNet E-serials Archive gopher://gopher.cic.net:2000/11/e-serials/archive||Unmanaged archive, managed collection in development|
|North Carolina State University, Electronic Texts, Journals, Newsletters, Magazines and Collections -- Scholarly Journals http://www.lib.ncsu.edu/stacks/stacks-Scholarly.html||Last updated in July|