Introduction

The topic of this Nature web focus, the future of access to the scientific literature, is the subject of lively debate among librarians, publishers, learned societies, and scientists. Much of the debate is about whether the literature should be 'open access,' and if so, how the costs of publishing should be met, and by whom.

Should scientific publishing, as some argue, be primarily financed upfront by funding agencies, akin to the way in which public funds are often used to pay for the publishing of genome, astronomical and crystallographic data in large databases? Or are the needs of scientists, quality, and innovation, best served by a competitive market in publishing? Or is the optimal solution a mix of both, depending of what it is that is being published?

This focus seeks to address such questions, and to distil the central elements of this often obscure debate for scientists at the bench. Lurking in the background is a more fundamental question; what scientists themselves see as their major needs in this electronic age. The journal itself is proving remarkably resilient as the main vehicle of scientific communication.

Three years ago, Nature ran a web focus on the same theme; "Future e-access to the primary literature?" It is notable that many of the opinions expressed then are still very relevant today, and that many of the issues concerning the primary literature (reports of original research) remain unresolved.

The last few months, however, have been eventful in moves to make the primary scientific literature available on open-access. Last October, the Public Library of Science (PLoS) launched PLoS Biology, an open-access journal which it hopes might compete with the few highest-tier journals (see Nature 425, 554–555; 2003). The UK's Wellcome Trust, one of the world's biggest medical charities, and some other funding agencies have also endorsed researchers using grants to cover open-access author charges.

PLoS's model relies on charging authors a $1,500 'dissemination' fee to cover the costs of publishing. Whether this form of open access is broadly sustainable has yet to be demonstrated, however. Indeed, a key question running through this Focus is whether the economics can be made to add up. Other questions abound: will scientists, or their host institutions and funding bodies, embrace the 'author-pays' model, and what are its implications, for example, on opportunities for publishing, and on the capital investments needed by any publisher to nurture new journals in their early years before they break-even.

One jarring aspect of proposals to reform scholarly publishing is that, all too often, they implicitly consider 'journals' as a single homogenous entity, to which one universal publishing model can be applied. On the contrary, diversity is everywhere. In any discipline, journals range from high quality 'must reads' with high rejection rates — which in turn result in higher costs per published paper — to publications which add little value to the articles as submitted, and are read by few apart from the authors themselves.

Journals are also published by a range of patrons, from individuals, and commercial publishers, to learned societies who use publication revenues to support their community in other ways. Likewise, a journal might be run largely by scientists working for free, or by professional editors. Some are electronic only, some have print editions. The list goes on. Any discussion of publishing models must surely take into account this heterogeneity. There is no one-size-fits-all solution.

Readers and authors appreciate the professional input that goes into producing readable journals of high scientific and editorial quality. And many of the challenges faced by the staff of a few journals such as Nature or Science are shared by editors of journals everywhere. But another questionable aspect of many discussions of science publishing is an assumption that the process of publication, as an endpoint of the research process, itself adds negligible value to the work. In practice, that added value varies greatly between journals too.

The core functions of publishing at its best share a commitment to impose intellectual rigour and high editorial standards on an exponentially increasing body of knowledge: to distil out the most important parts of that information, make it more accessible and place it in a wider context. As the flood of information grows, more and not less human editorial skill will be needed to focus and make sense of it, and the invisible hands of professional publishers and editors, with their associated costs, will if anything become more crucial.

Could these costs be paid for in other ways than the traditional reader-pays subscription model, and under what circumstances? Or will the outcome be a mix, with open access prevailing for certain types of publication, and elsewhere, market demand for greater access to the literature driving imaginative deals between publishers and libraries to make such access more affordable?

Last August, the Association of Learned and Professional Society Publishers declared itself "wholly in favour of maximizing access to research literature," adding that "the various proposals for achieving this... raise complex economic, logistical and sociological questions which differ from field to field as well as between different sizes and types of publishers. Much more information needs to be gathered through experimentation and analysis." That position still seems appropriate.

Declan Butler
European correspondent, Nature
Comments may be addressed to d.butler@nature.com