On Tue Mar 23, 2010 at 09:36:50 -0700, Michael Papet wrote: > Generally speaking, Bayesian filters seem to fail on > bodies with little text in them. Is that an accurate observation?
I think it is, but spamassassin (which you mentioned specifically) has a lot more testing options than mere Bayesian body analysis. Using my own filtering system, I've not noticed that short-body messages are the ones that get through more than anything else, so I guess I'm not the target market for you. Steve --
