Institutional repositories and the problem of versions

“One of the (many) problems I see our institutional repository running into with busy academics is the utter confusion about what they're allowed to upload into it. We promise them we'll check all the copyright for them (at Sherpa/Romeo) and won't put anything live that shouldn't be live. Which solves half the problem. But the other half of the problem is that even when the journal does allow some version to go up, it's always a different version. Some say the preprint is okay but nothing else; some say only the postprint; some say the final publisher's version. So when the author is dutifully filling out the details in the repository submission form, unless they go and do their research (something which we're telling them they don't have to do) they don't know what version to upload. Even if they do know the difference between a preprint and a postprint or that something labelled ‘author's copy’ is nevertheless the publisher's version. So, what we need is some magic DSpace (etc) plugin which, once the author's filled out the journal field, goes to look that up on Sherpa/Romeo and pops up a wee box that says ‘Cool, you can upload your preprint - by which we mean [insert clear and concise definition here].’ Of course I use the word ‘magic’ advisedly: from what I hear of DSpace this'd be difficult to impossible, and I don't imagine other repository management software is light years better. Ideas are easy, implementation is hard. (The third half of the problem is that it seems some academics don't actually save these previous versions. A solution to this is probably even harder to centrally automate.)”