Academics often use hand-rolled systems to keep track of and (less often, sadly) share literature references. I have used my personal wiki to that end for a while, but it wasn't the ideal solution.

Now, the rapidly-developing CiteULike looks quite interesting. It borrows from del.icio.us' simple interface and social software features, but it is tailor-made for academic papers that are available online. It lets you build a "personal library" (here's the one I just started), recording bibliographic information and enabling you to tag papers for future retrieval and group sharing. For instance, here is an ongoing stream of papers on blogging, collected by various individuals. Development is very much alive, as you can see from the development journal and the discussion list.

Because so much of the literature is still stuck behind subscription walls, surfing CiteULike can be frustrating if you're not on a university network, as you can very often be denied access to anything beyond the abstracts (even if you are, digital bouncers are legion and you're bound to bump into one of them sooner or later). This highlights how nice it would be for the public to have open access to the published research it has often paid for out of its own pocket. (The general web-unfriendliness of academic production is a pet peeve of mine - it hurts the impact and dissemination of research findings, and obviously deprives academia from influence on the "real world". How ironic that the Web was originally built in a research lab, to share results...)

(A similar service is Connotea, but I haven't done a thorough comparison between the two. And Alf Eaton's pioneering Biologging has been providing a similar service for biomedical researchers for a while now.)

(cross-posted to my personal weblog)