I'd like to compare two published versions rather than individual revisions for a single version. It can be hard to know what changes actually happened when you only fix a typo or two and then publish.