Making peer reviewers' comments public — not necessarily signed — would alleviate most of the problems outlined by Hidde Ploegh (Nature 472, 391; 2011).
Readers of the comments would then be able to judge, for example, whether reviewer requests for additional experiments were reasonable. Such a public-review policy would help editors and add a new dimension to a journal's reputation, particularly if others in the field publicly shared their own relevant observations.
In conventional peer review, especially at top-tier journals, much of the reviewing effort goes into manuscripts that are ultimately rejected, meaning that the scientific community has no access to these communications. Under a public system, these records could prevent reinventions of the wheel and help educate newcomers to the field or to peer reviewing.
Publicly available reviews, including those of rejected manuscripts, would also provide an incentive for authors to submit their manuscript only when it is ready — helping to lower rejection rates and aiding the search for suitable reviewers (see http://go.nature.com/qamrfc).