John Ewing, in Correspondence, argues that open-access journals are not open to everyone because not all authors can pay or find a sponsor to pay their processing fees (Nature 425, 559; 200310.1038/425559a). Although publishers of open-access journals such as the Public Library of Science (PloS) say that authors who can't pay won't have to, Ewing feels they have underestimated the numbers who will not be able to pay.
As a long-time advocate of open access to science, I can list several points to suggest that this concern is overstated.
Declan Butler notes in his News Feature “Scientific publishing: who will pay for open access?” (Nature 425, 554–555; 2003) that many funding organizations are willing to pay these fees for their grant recipients. Although this solution will not work in disciplines that are less well-funded than medicine (in Ewing's field of mathematics, for example, or mine, philosophy), that is no objection to applying it to fields where it can work. It is highly likely, and desirable, that different fields will develop different open-access models. Some peer-reviewed open-access journals in the humanities charge no processing fees at all.
Ewing notes that many universities may not be able to afford these fees, in cases when funding bodies cannot pay them. But if open access spreads, every university will make large savings from the conversion, cancellation or demise of expensive subscription-based journals. This money can be used to support the open-access model of archiving and publication.
If some future open-access publishers have no policy to waive fees, and authors find themselves excluded on financial grounds alone, there are other ways to bring about open access to peer-reviewed literature (such as e-print archiving) that do not depend on open-access journals, research grants, affluent employers or windfall savings in the library budget.