My 2 cents:
- I would really like to see the reviews published next to the paper. This way the work done by the reviewers would not be lost to the community, the readers would be able to see which points were contentious or not, the reader could also see whether a given methodological point (dear to the readers heart let's assume) was covered, everyone could see how much the review process improves many papers submitted to PLOS One (in some cases I feel the reviewers contributed more than some authors). I fail to see any drawbacks whatsover.
- BUT I don't believe in publishing the reviewers' names, at least in the present state of things. Too many possibilities of pressure, retaliation, etc.
- I am also favorable to double-blind. I've read that in quite a few cases the reviewer can guess the authors. I fail to see how it's a problem. Similar to publishing the reviews, I don't see any potential drawbacks.