John Johnson wrote:
The problem with spam is not the content, but the conversation.
No, I disagree. I perceive the content to be spam, much
in thw way tools like SpamAssasin and Vipul's Razor judge
the *content* in an automated fashion. Or, an ISP
client. They could care less about the "envelope".
They care that the content is objectionable.
SpamAssassin and Vipul's Razor don't work that way at all.
SpamAssassin works by scoring messages based upon how "spammy" a message
looks. Not "content" per-se, but more along the lines of "things that
look like spammer behaviour" (ie: fingerprints of extreme hard sell
behaviour mostly unique to spammers, forged headers, fingerprints in
envelopes, appearance in blacklists (if feature turned on) etc.).
Vipul's Razor, on the other hand, works more along the ways of "this is
extremely similar to what we've previously judged as being spam" - this
is more of an identification of the sender and his pitch, not the pitch
"instance" itself (because a different sender would be using a different
pitch).
In neither case are they trying to intuit what the content _is_, but
rather an approximation of either "this looks like spammer tricks" or
"we've seen this one before" respectively.
While SpamAssassin can be forced into looking for generic "types" of
content, that's not its intent or default configuration.
Spamming is a behaviour, not a content. As an extreme example, a
"consumer" of porn could very well be quite happy to receive _solicited_
advertisements from a porn site, but be quite adamant about refusing
unsolicited ones.
_______________________________________________
Asrg mailing list
Asrg(_at_)ietf(_dot_)org
https://www1.ietf.org/mailman/listinfo/asrg