I'm sure that the author of "A plan for spam" (introducing the idea of bayesian filtering to recognize spam) can find a way to classify nasty comments.