If they have "foolishly" already signed that right away to a journal for a final draft accepted for publication, he says, then they should simply self-archive instead a penultimate draft, appending a list of any errata and updates.

BBC: Setting research papers free