FWIW, MailScanner (http://mailscanner.info) which is a server-side tool (it
is very useful if you don't need to personalize settings or rules per user),
has an option which allows you to have 2 sets of SpamAssassin rules.
The standard set does its antispam thing.
The second set is geared toward this kind of things (you still have to write
the rules by yourself).
The good thing is that you can decide what you do with a message, based on
spam scoring and on this other content scoring.
For example, you could deliver spam with a mark, but delete the message if
the 'content' score is too high.
Independently of the tool, take into account that effective censorship is
really hard to do, and if you are not extremely careful, you may very well be
banning completely innocent content.
I've once come across a rule that hit the word 'cum' but did not check for
boundaries, so 'document' would hit...
Regards.
El 17 Jun 2004 a las 0:00, Matthias Fuhrmann escribi�:
> On Wed, 16 Jun 2004, Jeff Koch wrote:
>
> >
> > We need to come up with an approach to provide content censoring for some
> > schools and religious institutions that we work with. Basically it would
> > involve quarantining mail that contained certain 'bad' words. Could SA be
> > used as a 'bad' word content scorer? Perhaps we could add some special
> > rules to look for these words and score the email accordingly. Has anyone
> > already done this?
> build some rules like this:
>
> body BAD_WORD_RULE /(bad|word|and|other|bad|things)/
> descibe BAD_WORD_RULE some bad words
> score BAD_WORD_RULE 100
>
> regards,
> Matthias
--
Mariano Absatz
El Baby
----------------------------------------------------------
Unix is very simple, but it takes a genius to understand the simplicity.
-- Dennis Ritchie
