Harvard is apparently adopting an opt-out OA policy. This mechanism of encouraging OA is brilliant and entirely in line with the research on opt-out in organ donation and IRA adoption. Because there's an opt-out provision, no one can complain that this restricts their journal choices. Because it's there by default, adoption will increase dramatically. The only issue is that I imagine many people will sign traditional journal agreements anyway, and thus wind up placing themselves in legal messes. Not that big a problem though, because journals would be even more misguided than the RIAA to sue. Think that suing customers is bad? Try suing academics for disseminating their works.

Ari