Reading your comment, one that entire agree with, I realize that the «review» is pretty loaded, putting the «reviewer» in a place of authority by default.

While most of the time, the review occurs when work seemed good enough for the requirements, the effort and cost in time in accord with real world micro economics, meaning that code is submitted by a professional doing the best effort in his judgment, and you better have good reasons for your feedback.

In that light, it's pretty clear why things can go sour pretty quickly if a reviewer is not cautious with his feedback, and also why it's hard to make it work if reviews are enforced.