Prompted by Peter Suber's blog, I disvoered a College and Research Libraries article by David W. Lewis, dean of the Indiana University-Purdue University Indianapolis Library, which proposes a strategy for moving library funds from journal purchase to OA support and preservation of the book-buying budget. The logic is inescapable - university libraries are even less able to continue to put up with journal price increases than they were (and if that is a problem in the USA, what is it like anywhere else?) and with repositories growing in number and in number of items held, the grip of the commercial publisher cannot be maintained.
However, the article doesn't really go far enough. Action by libraries without further action by the parent institutions will not resolve the issue: if OA is regarded as the "Gold" of author charging, the institution, in one way or another will have to fund the submission of papers. The only sensible strategy is collaborative, subsidised, freely available OA publishing. Imagine a journal published by a consortium of major US universities - some of the Carnegie Research I group, for example, which covers all the major research institutions - it is likely that such a journal would be rapidly accepted as a major outlet for research and if those universities required that faculty offer their papers first to that journal, it would soon become essential reading in its discipline and it would rapidly shift awareness of the value of OA. Tinkering around with the present system through repositories and author charging will never bring about the necessary changes to make access to scholarly research available freely to all. A much more radical break with the past is necessary.