Jim Till has posted a comment on the IRCSET policy.
Peter Suber and I have speculated that the IRCSET policy might be the best model so far, however Jim points out that the combination of the first key principle: 1. This publication policy confirms the freedom of researchers to publish first wherever they feel is the most appropriate, and the stipulation "should adhere", might constitute a loophole which would render this policy ineffective.
Good points, Jim!
Jim Till, author of Be Openly Accessible or Be Obscure, is one of the reasons we have Canadian Leadership in the Open Access Movement.