I have enjoyed the discussion in the various biblioblogs on Open Access and the need for a stable repository for articles. If you've not been following it, the best place to go is Hypotyposeis:
Open-Access Needs a Stable Repository
This keeps up to date with the links on the different biblioblogs so that you can thread the discussion. It's interesting in particular to hear that Deinde have some plans. I don't have any strong thoughts myself about the topic except to agree with Tim Bulkeley's tentative suggestion that SBL would be a good place for this, if we could persuade them to take this seriously.
My other comment would relate to the issue of hosting articles (repository proper) vs. linking to articles (Gateway). The word "repository" often goes with "open access" and not surprisingly so. But whose articles are going into the repository? Is it not effectively those least committed to open access? What I mean is that those scholars already committed to open access are those whose articles are already on the web, made available via their homepages. How are we going to persuade the others, who have not so far found it in them to publish to the web, to publish to the on-line repository? Is there a mismatch here? I hope not. I am strongly in favour of open access. I just wonder whether the ultimate model is the one already in procgress, the evolutionary one, in which scholars are steadily becoming convinced of the ideal and uploading to their own web space.