During the six years that I was the editor of the Journal of the Medical Library Association it was rare in the extreme that we published an article exactly as submitted. When I started, I suppose I thought of peer review as strictly a search for error, an opinion as to whether or not the piece should be published. I came to understand that the reality is far more subtle and complex.
Marcus suggests that librarian journals should evolve into blogs, arguing, in part, that peer review should be a post-publication process rather than a "pre-publication process that sometimes drags out for many months." Although there is something appealing about this idea, when I think about the actual articles that I was involved in editing, I'm not at all sure that this would be a good thing.
As editor, I considered my responsibilities to be not just to select the most appropriate work for each issue of the JMLA, but to work with each author to present their work and ideas as successfully as possible. And I began to see the peer review process as just the first step. Often (perhaps usually) the reviewers (we always had three reviews) disagreed among themselves as to what the major issues or necessary revisions were. Sometimes they flat out contradicted each other, and part of my role was to sort through those differences and give the author the best advice I could. Frequently, the writing was simply poor and there were many cases in which it became clear to me as I worked my way through the article that the paragraph that the author had written actually said nearly the opposite of what they intended to say. Or the article would be full of extraneous and repetitive material that simply got in the way of a busy reader actually getting to the meat of the work. I believe that most of the articles that we published were more effective in connecting with an audience than they would have been without the pre-publication review and editing.
I'm not at all sure that it would be a service to the library community if all of those articles that I read through in their first iterations had simply been posted to a blog and opened up for comment. The few experiments that have been done in the last couple of years with post-publication review have not been overwhelmingly successful, the ArXiv experience notwithstanding (extrapolating from the experience of a small, tightly knit, fairly homogeneous scientific community that writes largely in formulas to a much more diverse, narrative literature is a stretch, I think).
Certainly, the lag between completion of an article and its availability to a wide audience needs to be shortened, but this is a matter of efficiency, not a fundamental aspect of "traditional" journal publishing. The major scientific journals have typically reduced this time lag to a few weeks and there's no structural reason that can't be done with the library literature -- it's a matter of the resources that you're willing to devote to tackling the problem. Marcus suggests that this could be accomplished by "a carefully designed and managed blog." The devil is in the details, and I wonder if that careful design and management would result in something that looks more like a typical journal or more like a typical blog.
I'm not one who is terribly impressed by the "wisdom of crowds" (a concept that seems to be especially dubious during the US election season). I've rarely seen anything approaching substantive discussion and analysis take place in a comment thread, and the longer the thread, the more worthless it typically is. Rather than providing vibrant post-publication review, I'm afraid that posting unedited articles for comment would result in much good work being buried and ignored.
But the terrain continues to evolve rapidly, and the opposition of blogs to traditional journals is probably a false distinction. The traditional journal is rapidly morphing into something else, while adopting features that we associate with blogs (the ability to provide rapid responses being the most obvious). The underlying architecture of blog software is also developing rapidly and trying to accurately characterize the qualities that define a blog would likely become a contentious debate.
Marcus is pushing the right questions, and everyone involved in scholarly publishing, at whatever level, should be thinking creatively about how to make the communication and discussion of projects and ideas more effective. But it isn't a matter of journals vs blogs. The most effective modes of communication that we develop over the next decade will adopt features that we associate with each, but will be fundamentally different from either.