In connection with yesterday's story about a publisher trying to stifle fair use rights on a journal article, I learned today of another interesting development in open access publishing: The UC system is proposing a system of forced open access for all UC faculty publications. This appears to mean that
(1) The UC system would negotiate collectively with publishers to allow all journal and conference articles to be released to an open-access repository such as arxiv.
(2) Open-access locations such as arxiv ids would be included in lists of publications as a standard part of the tenure and promotion process.
(3) Modulo some kind of bureaucratic exception process, and despite continuing to hold copyright in their own works, UC faculty would automatically extend to the UC Regents the right to republish all their journal and conference papers in open-access repositories.
I haven't had a chance to talk to very many other UC faculty about this, but obviously (3) is likely to be a lot more controversial than (1) and (2)... Though I wouldn't so much mind myself. I and I think many of my colleagues would be happy to have our papers as widely disseminated as possible, and a step like this seems necessary in order to make my lazier colleagues' work available.
Publishing a paper only in X random journal is, in the long term, useless. It might appeal to your peers in a limited way, and you perhaps want to gain extra income from its publication, but it puts you into an insignificant rathole no one will ever read.
If you're, really, really lucky in the way that the very random musician is who gets a record contract after selling out to a producer, you will get a boost from a publication, but if no one can read it, you vanish into obscurity with the rest of your nameless compatriots.
I applaud the university in making it a requirement that your publications be freely released. If you don't do it, you don't really understand science or the social contract of academia. That makes you a pointless self-aggrandizing attention seeker, and probably one who has no real contributions to make.
The above shold not be read as aimed at the poster, who appears both open and sensitive to the academic process, but rather at those who would find the open publication controversial.
Most computer scientists I know of agree, and generally put all their papers online somehow (if only on their own web sites) for exactly this reason. Sites such as arxiv have the added advantage compared to personal sites that they have people paying attention to making sure the papers don't get lost in disk crashes or hardware upgrades or post-mortem or whatever.
After posting this, I went to a campus faculty forum about the policy. The biggest concern I heard, from faculty in less tech-savvy areas, was that somehow this sort of publishing would put the journals out of business (despite zero evidence from physics and CS, where online versions are common, that it would hurt the journals at all). The journals do still serve a useful purpose in maintaining quality control so that one can find the good papers among all the junk, but I think that purpose is served better by putting their good names on papers than by making the papers available for a price.