(2007) Moderation isn’t rocket science

Posted again here despite its great age because "Moderate by hand versus by statistical inference" debates go back to at least the invention of NNTP[0] but flare up here and on lobste.rs/reddit every September/December/January.

0: See Bellovin, et al, in Steve Bellovin's USENIX award acceptance materials. https://www.usenix.org/about/awards/flame and https://www.usenix.org/system/files/login/articles/07_bellov... for those who are not familiar with USENIX and USENIX's Security conferences.