[REBOL] Re: Cerebrus 1.0.5
From: gchiu:compkarori at: 20-Nov-2002 21:09
On Wed, 20 Nov 2002 16:56:04 +1300
"Andrew Martin" <[Al--Bri--xtra--co--nz]> wrote:
>So far it's working OK. It sometimes doesn't classify
>some spam email as
>spam email, but I think that's because these emails
>haven't been reported to
>spam cop or similar.
Hi Andrew,
I think what happens is that the black listing DNS
services like SpamCop run multiple spamtrap addresses, and
when these addresses get mail, they report in to SpamCop
who then tallies a score and marks a mail server as being
a spam source when it is over a critical threshold.
So, it depends upon which lists these spamtrap addresses
have been harvested into, and if your email address is not
in the same database as theirs, then it could be they
don't see the spam you get.
Some anti-spam programs allow the users to form a
collaborative network where a user reports an instance of
spam, and the IP address of the mail server where that
spam came from is then notified somehow to all the other
users of that particular program.
I was thinking that Cerebrus could do the same. Perhaps,
a rugby server that accepts submissions from trusted
clients ( to prevent the spammers from inputting false
data ), and another rubgy server that allows Cerebrus
clients to query. I guess it would be faster than doing
it as cgi, but I risk my ISP killing my ruby service as it
would be a persistent process which is not allowed under
TOS.
--
Graham Chiu
