Publishers' scepticism about the claims of the open-access campaigners ("Everyone's a winner", May 7) is easily caricatured as self-interest. But it goes deeper.
The Wellcome Trust's estimate of author-pays publishing costs is inadequate. Like subscription journals, author-pays journals have sales forces that negotiate with authors and institutions to pay their fees. The system will also need elaborate invoicing and management systems to track and collect submission and publication charges from a globally dispersed research community. The costs could end up as high as - and possibly higher than - that of subscription journals.
And institutions will incur new costs under such a system to cover the authorisation and administration of payments to publishers.
The Wellcome Trust estimates, although optimistically low, are higher than the amount that authors are prepared to pay - 50 per cent are not prepared to pay anything, while 80 per cent of the remainder are not willing to pay more than $500 (£280) per article, according to a recent international study.
Furthermore, the Wellcome estimates are considerably higher than the sum charged by any author-pays publisher at present. It seems that cross-subsidies will be necessary to cover the shortfall - but how long can this model be sustained?
So the Wellcome Trust proposes a different way of squaring the circle: authors should pay to submit their papers for review. Under this system, rejected authors will be expected to pay most of the publishing costs and successful submissions will pay again.
But if authors are reluctant to pay for publication, why would they agree to pay for rejection? This is not so much an author-pays model, as a rejected author pays - and it will create a universal financial barrier for prospective authors, one that does not exist today.