On Mon, Aug 30, 2010 at 02:03, RW <rwmailli...@googlemail.com> wrote: > On Sun, 29 Aug 2010 17:36:36 -0700 (PDT) > joker_ft <gugafontan...@gmail.com> top-posted: > >> Benny Pedersen wrote: >> > >> > On søn 29 aug 2010 17:28:52 CEST, joker_ft wrote >> > >> >> Does anyone know some public corpus updates in 2010 ? or why the >> >> spam assassin public corpus stopped update in 2006 ? >> > >> > if spammers know what being scanned for will it be effitive >> > stopping spam ? >> > > >> >> Yes, this make sense. > > > No it doesn't. If spammers want to know what SA is looking for they > can just download it. If they want to optimize their spam to resist > Bayes there's nothing special about corpora used to develop it.
Exactly -- nothing about SA's ruleset requires that the coding or corpora be kept secret. The only reason we keep most of our corpora private is due to the contents being mostly private mail. We don't have time to update the public corpus, unfortunately; it's quite labour-intensive. --j.
