A hybrid model with both assigned reviewers and public comment is being tried by a regional journal (Southern Association for Information Systems): http://knol.google.com/k/craig-vanslyke/jsais-review-process...