Re: automated spam detection

[Richard Kettlewell <]

G.W. Mills Beebit writes:
> Richard Kettlewell <richard.kettlewell@localhost wrote:

>> These thoughts derive from some recent discussion I spotted on
>> Usenet...
>> 
> <scheme for "logging" MD5 hashes with "central" server>
>> 
>> What are the potential problems of this approach?
>> 
> What about _legitimate_ (opt-in) mailing list traffic?

That's a good point about any kind of scheme which aims to detect bulk
mail.

As far as I can see the only practical different between legitimate
traffic and spam is that the recipient has opted to receive it;
probably it's at the point where they make that decision that the spam
detector has to be updated.

That's not going to be practical any time soon, given the extremely
diverse set of ways that people register for mailing lists.

ttfn/rjk