I like it when journals publish articles about academic publishing. Now, this one, called “A Good Review”, doesn’t say too much, but it does have a nice list of dot points: (and I quote)
A good review helps the members of the scientific community achieve standards higher than what they might be able to do without expert feedback.
A good review helps the authors learn something new or consider something they had not thought about.
A good review helps to improve the communication of the material and alerts the authors on statements that may be misleading, misunderstood or plain wrong.
A good review is done in good faith; it addresses the contents of the manuscript at hand not the state, status or character of the authors.
A good review is not about the expertise or cleverness of the reviewer, it is about the quality of the proffered manuscript—and, really, nothing else.
Me again: Generally, in my limited experience, I would say that feedback from the reviewing process has made my (few) papers better. At times, you don’t really want to hear what they have to say, but after changing things around and spending some extra time, the manuscript is improved.
Having said that, I don’t believe the huge lag times for this process are justified, and we’d all be better off with a more informal system like arXiv. The papers that are good will still be cited and read. Despite the whole review process, you still get some stinkers even in the so-called “high impact” journals.
Rather than hindering the publication of new material, we “just” need a better way to catalogue and access what’s already there. Should new postgrads really have to re-create entire literature surveys for every single project? One day, I would like to create the “Wikipedia of literature reviews”. But not like Wikipedia, coz that’s not the best model for this sort of information.