Mathematical Reviews is a publication of the American Mathematics Society that provides capsule reviews of essentially every mathematics research paper that is published. It has been in operation since 1940, and in recent years the entire database has been on-line, with the name MathSciNet. Some of the reviews are rather entertaining in one way or another, and the availability of this database has led to a new form of diversion among mathematicians: searching on promising words (e.g. “plagiarism”) to see what turns up. My colleague Kimball Martin has put together a new web-page entitled Exceptional MathReviews with links to the best reviews he has heard about.
One of the most famous such reviews turns out to be apocryphal. In the days before the searchable database, I had heard it claimed that one Math Review contained the following devastating evaluation: “This paper fills a much-needed gap in the literature.” It turns out that this isn’t actually true. For the real story and more about Math Reviews, see an article from the 1997 Notices of the AMS.