Nature news titled recently that "Rebel academics ponder how to break free of commercial publishers". The rebels would be better off if they'd read this blog, because we have discussed here a solution to their problem!
The solution is Pre-Print Peer Review (PPPR). The idea is a simple as obvious: Scientists and publishers likewise would benefit if we'd just disentangle the quality assessment from the selection for journal publication. There is no reason why peer review should be tied to the publishing process, so don't. Instead, create independent institutions (ideally several) that mediate peer review. These institutions may be run by scientific publishers. In fact, that would be the easiest and fastest way to do it, and the way most likely to succeed because the infrastructure and expertise is already in place.
The advantages of PPPR over the present system are: There is no more loss of time (and thereby cost) by repeated reviews in different journals. Reports could be used with non-peer-reviewed open access databases, or with grant applications.
Editors of scientific journals could still decide for themselves if they want to follow the advice of these reports. Initially, it is likely they will be skeptical and insist on further reports. The hope is that over time, PPPR would gain trust, and the reports would become more widely accepted.
In contrast to more radical options, PPPR has a good chance of success because it is very close to the present system and would work very similar. And it is of advantage for everybody involved.
I have a longer outline of the idea here, comments and suggestions are very welcome!