Recent discussions of the government-sponsored Working Group on Expanding Access to Published Research Findings rightly stress its somewhat limited vision of the "gold" open-access model, pointing out that gold merely shifts the cost burden of publishing from readers to authors and, through them, to research funders and universities (eg, "Don't deal in a debased currency - go green", Opinion, 5 July). The Finch group thus keeps intact the profits that pre-digital age publishers have been able to collect, while adding an additional ?60 million to pay for the article-processing charges levied by such commercial outfits.

While such a strategy supports the rights of readers freely to access research literature through either the gold or more supportable "green" (repository-based) model, it effectively ignores the rights of authors freely to publish their research. The Finch report contains no proper discussion of who gains and loses under the existing arrangements, and how this might change under the charging system it favours. Without such analysis (including evaluating the role of academic reviewers and editors who supply their services for free), it is difficult to come to any rational conclusion about the open-access model we need.

Academics without access to funding sources for article-processing charges, especially those in developing countries, would have little option but to rely on publishers' altruism or meet the costs themselves. It is interesting to note that the Finch group's basic terms of reference were to look at "expanding access to the published body of research", but with no reference to the wider issue of how a publication system could be devised to promote high-quality research in an optimal and equitable manner.

Remarkably, the Finch report makes no mention of the alternative strategy of direct support for journals founded on the principles of true open access. Unlike the huge charges levied by existing commercial publishers (especially those entrepreneurs who have recently entered the open-access market), such journals seek to recover no more than their break-even running costs, usually through society membership or nominal/voluntary reader registration fees.

ÁñÁ«ÊÓƵ

Fortunately, the availability of high-quality open source software has encouraged the emergence of these open-access journals, giving them the chance to demonstrate the superiority of their model.

Unfortunately, a golden opportunity to transform an expensive and antiquated system into one fit for the 21st century appears to have been missed, at least for now.

ÁñÁ«ÊÓƵ

John Bynner, Emeritus professor of social sciences in education Institute of Education

Harvey Goldstein, Professor of social statistics, University of Bristol