Department of Electronics and Computer Science, Southampton University
Posted to the Web: January 1995; last modified November 1997
**See author's note added November 1997**
Principally, electronic publications, where they exist, do not fulfil the academic ‘reward’ structure, and authors justifiably recognize that they still need a good publisher’s marque on their work. And for conventional print-based publishers there are at present few incentives to promote the transition to formats that are electronic from inception.
To examine why such a transition is achievable and why it may be desirable for academic authors, and to understand why many publishers may resist change on this scale, it is worth looking first at a model of the publishing process. This analysis suggests possible structures upon which the change to electronic publishing can be supported, and considers the prospects for Harnad’s ‘subversive’ proposal  - that individual ftp archives be established by every academic author and made available free to those that wish to read them - the aim of which is to motivate the change from within the community if the change is not forthcoming outside.
Although the technological framework that would support electronic journals is operational, and continues to develop at some pace, it is the social and economic impacts that may prove more intractable. The paper points to research that will link present and future technological capabilities with new publishing structures.
A simple model of the traditional journal publishing process (Fig. 1) shows the development stages of the fundamental unit of published work, the paper. The publisher is heavily involved throughout, both directly and indirectly, through the Editor, often an academic but who is responsible to the publisher. Depending on the nature of the journal, papers may be commissioned and directed, to a greater or lesser extent, extending the publisher’s influence right into the authoring process. Clearly, the publisher takes complete responsibility for the collected work, the journal issue.
Fig. 1 The traditional publishing process
The advent of desk-top publishing (DTP) tools widened the scope for authors to be involved in the production of their papers, up to the point of producing almost typeset-quality copy that could be used to print directly from disc or paper if desired. DTP tools were ideal for the production of preprint conference proceedings, for example, but such publications do not carry the hallmark of an established journal. In many cases it proved difficult to integrate papers produced using different word processing or DTP formats with existing journal styles, and journal publishers preferred to stick with established procedures.
But authors’ growing experience of the ’typesetting’ process and consequent desire to control the appearance of their work has since been encouraged by publishers anxious to cut costs and maintain profit margins in the face of fragmenting markets, squeezed library budgets and declining sales. and it is possible to use the author’s generated copy.
So the influence of the author has grown to encompass the whole production process for the paper in some cases (Fig. 2), but the publisher is still responsible for the core functions of marketing and distribution - selling the product into the market - on which any good publisher can justify its role.
Fig.2 The modern publishing process (possible) - the diminishing role of the publisher
Even in these areas, however, academics were becoming more sophisticated, creating special-interest groups, many international, in the increasingly inter-disciplinary fields that were emerging, often in the absence of immediate support from a professional institution or a recognized publication. Such groups, of course, usually initiated and created new publications with publishers, especially journals. In most cases their ‘informal’ networks became the basis of publishers’ marketing efforts for the new journals, indicating just how important and wide-ranging the academics’ role has become. Academics’ experience with conference and course organization - including the marketing of conferences as well as the production of proceedings referred to above, and wider, more proactive promotion of university courses - should not be underestimated in this respect.
Informal networks have proved extremely effective for specialized fields, but stronger support was still necessary for publications with expanding markets as it became either infeasible or undesirable for academic groups to manage the complexity of growing publications.
From the viewpoint of scholarly journals the Internet is the ideal distribution medium. Computers used to prepare, often complex, copy are invariably linked to this worldwide network, and material can, in principle, be disseminated at the press of a button. A tool that potentially both reduces the complexity of publishing a large body of material and provides an inherent distribution mechanism has emerged. It is even possible to argue that the appearance of search tools such as on the World Wide Web, itself developed by academics, provides a pseudo-marketing function, among other functions, that will assist publishing applications. These 'Net' tools complete a scenario in which the authors--editors--readers of a scholarly publication could take complete control of the publishing process (Fig. 3 ).
Fig. 3 The publishing process for the electronic journal
The academic community - foremost, authors and readers - has to want
the change if it is to take place. On the cost issue alone, that is, the
cost of the publications, there would be little demand for change. The
publishing process is expensive, as reflected in journal prices, but it
works, it serves the academic community effectively and is firmly embedded
in the academic reward structure. The main resistance to the cost structure
is currently not from authors and readers, but from libraries.
Publishers have a number of weapons to protect their interests (Fig. 5). The most effective will be the ownership of publications with strong identities established over many years through effective management and marketing. The academic reward structure is predicated on such titles. Highly motivated scholars can transfer the interests of a given field to the new medium, but not established titles without the involvement of the publisher.
This, in turn, is the first dilemma for users of the new medium: is it simply to become an outlet for material from the weaker journals, while the best work is forced to publish in the recognized paper journals? This need not be so if the uniqueness of the new medium is exploited in the development of new titles.
The cost implications now need to be addressed. If publishers see no commercial advantage - reduced profits in absolute if not percentage terms - they will not support a new electronic publication. Harnad  argued that the cost of a purely electronic journal will be a fraction - he estimates 25% - of the equivalent paper journal. Given the progressive transfer of control illustrated in Figs 1-3, this may indeed be the magnitude of the reduction in costs - typesetting, paper, printing and some marketing and distribution costs all disappear from the equation, as well as a significant part of the publishing overhead. But a significant residual cost remains and must be covered. At present there is no funding mechanism that will cover it. So what funding mechanism will work in this situation, and will governments and industrial sponsors support it? Cost savings there may be, but there is already evidence that this will be a political minefield.
Fig. 6 Publishers of the electronic journal
Harnad's 'subversive' proposal is clearly the most extreme, but it is a means to an end and is not a long-term solution by itself, although it could be an important precursor to other publishing structures.
What are the requirements a new publisher of scholarly electronic journals must satisfy? It must have close links with the academic community and it must be able to establish very quickly academic recognition of its publications. It must have experience in applying the supporting technology competently and reliably - the journals must work first time - and be professional, well organized and flexible enough to accommodate continuing developments. Importantly, the organizational structure must be consistent with the lower overheads that will prevail. Universities, academic groups, special-interest groups (SIGs) and learned societies are all organizations that could support these requirements - when a funding structure is identified.
Ginsparg’s physics archive  is the preeminent demonstration that this approach works, but it remains a preprint archive and thus does not convey full academic recognition of the papers it carries, even though the popularity and scope of the archive mean that it is far more widely used than most paper physics journals .
This highlights the other practical problem (Fig. 5), that of establishing a peer review structure in the electronic medium . In principle, establishing this process is as straightforward as building a review structure for a new journal, but it is new formats and not necessarily new fields that are being created and there will be hostility from those committed to ‘paper’ publishers and difficulties for those loyal even to looser ties with existing publishers. After all, the present scholarly publishing system is inimical in that that the key people in a given field will already be linked with existing publications serving that field.
The role of the learned society is interesting. The societies' publishing arms are invariably run on a commercial basis, which potentially puts them in conflict, as explained above, with the aims of publishing scholarly journals electronically. However, the members’ interests at large, probably in support of electronic publications, could be a strong influence. The American Physical Society and the Association of Computing Machinery, both also major publishers, have outlined strategies for creating electronic journal databases , .
Large publishers may also want a role. Those corporations with interests particularly in the illustrated non-fiction markets are already developing expertise in electronic publications. The restructuring needed to accommodate the lower cost structures in academic publishing will not be easy, but could be achieved where publishers decide that they must build a long-term interest in electronic fields and identify those projects where there is room for the publishers’ traditional ‘value adding’.
Dynamic reconfiguration of publishing groups within these corporations should at least be feasible. Such companies have largely divested themselves of non-publishing interests, Reed International being a prime example, so that they can be more responsive; flexibility and focus were two key arguments for Pearson’s recent dispersal of its Longman academic publishing group.
Software publishers are major producers of electronic publications, particularly in the multimedia CD-ROM form, and have spawned a number of start-up companies in this market, but few such publishers have interests in the academic journal market at present.