At 6:26 -0700 2003/05/13, Marc Perkel wrote:
>I think I'm going to use this and try it out. Except that I'm going to
>use ACCEPT because these lists are often not accurate.
what about using these lists to increase the score of an smtp transaction?
At the moment I am only using two of these with a certain weight in
the transaction score (I have both a rcpt and a whole transaction
score), but I will soon add the riskier ones with a lower weight.
Since I mentioned it, the score is actually based on a N-dimensional flag, the
flag then will have a certain distribution in an N-dimensional space,
it would be interesting to analyse the distribution and find an
optimal (perhaps non-linear) weight function to minimise the false
positives.
Am I that serious?
Giuliano
>Konrad Michels wrote:
>
>>You need something akin to this - although the delay bit is optional . .
>>.:
>>
>>deny message = rejected because $sender_host_address is in a
>>blacklist at $dnslist_domain\n$dnslist_text
>> dnslists = sbl.spamhaus.org=127.0.0.2
>> delay = 4m
>>
>>deny message = rejected because $sender_host_address is in a
>>blacklist at $dnslist_domain\n$dnslist_text
>> dnslists = opm.blitzed.org=127.0.0.2
> > delay = 4m
[...]
