As a supplement to other useful answers and comments: as I've said a number of times on these sites, the apparently immediate idea of open-ness has some issues, the idea of direct communication with authors of submissions one is refereeing has issues, and the idea that there is no impulse to game the system is pathologically naive. Think of retaliation on whistle-blowers, retaliation on sexual harassment complainants, and so on. My point is that it is injudicious to operate as though human beings would be dispassionate and rational... even while, or perhaps "ironically while"... we're talking about pseudo-rational issues such as scholarly merit.
When I referee papers (ok, math is my biz), I try to be positive, especially for junior people who need to get tenure ... or a job, but I do often also suggest substantive changes, and point out substantive infelicities. By this point in my life, maybe I'd be willing to bear a sort of public flak about my critiques/recommendations/edits, but I'd really rather not. So, if a publishing entity cannot commit to my anonymity as reviewer/referee/critic, I'd probably demur.
Lest people think I'm just being a baby, a coward, etc., I'd note that a few years ago I tried to help some good, young people edit their paper so that it was ... um... not literally fallacious. One of those situations where no one doubts the conclusion, but, ... srsly... the proofs should be genuine. ("Or is it just me?!?" ...) The authors did not understand the issue, got angry, etc. Now, while I am disappointed that they did not take my advice, or understand it, ... especially given that misunderstanding, I would not want to be known as the person who (to their minds) "did not understand their paper, and was a b*tch"...
That is, I don't mind giving other people the option to misunderstand what I'm saying, but I'd prefer to be out of the "sights" of their unhappiness.