This is an important discussion that has been going on in disparate places, but primarily at the moment is on the Open Science mailing list maintained by the OKF (see here for an archive of the relevant thread). To try and keep things together and because Yishay Mor asked, I thought I would try to summarize the current state of the debate.