[Date Prev][Date Next] [Thread Prev][Thread Next] [Date Index] [Thread Index]


On Fri, Oct 18, 2002 at 09:27:50PM -0400, Clint Adams wrote:
> You might want to consider stripping the X-Spam-Status headers or
> putting spamassassin -d in the pipeline before bogofilter, since your
> pristine input won't necessarily be tagged as such.

yep, i've been doing that - stripping headers with formail

it's working even better than i expected.

it's about as good at detecting spam as my old body_checks rules or
amavisd + spamasssassin was, but with a lower false positive rate.
haven't had any false negatives yet.

i'm quite impressed by bogofilter.  it's good software.  it's not the
whole solution, but it's a very important part of it.

> > after a few months, it may even be possible to take spamassassin out
> > of the procedure as bogofilter may be "trained" well enough to
> > identify spam without any help from SA.
> I got to this point well before a few months.

yeah, less than a day.

i'm not using it by itself, though.  i'm still using SA, mostly because
it's easier for me to directly influence what SA detects as spam, just
by adding rules to local.cf.  incoming email is piped through
bogofilter, and then into SA before delivery.  SA is configured to score
"X-Bogosity: Yes" as worth +10.0 points.  my SA threshold is 10.1

i.e. anything bogofilter says is spam is (almost certainly) actually
spam.  any spam that makes it past bogofilter, SA will probably detect
(and feed back into "bogofilter -S" so that bf gets trained some more).


craig sanders <cas@taz.net.au>

Fabricati Diem, PVNC.
 -- motto of the Ankh-Morpork City Watch

Reply to: