16 Jul
2002
16 Jul
'02
5:59 p.m.
"JWB" == John W Baxter <jwblist@olympus.net> writes:
JWB> I don't think the ISPs *can* solve the problem in the
JWB> near-to-medium-term future. [Longer term, with the demise of
JWB> SMTP and its "everything open to all except for a few
JWB> bandaids" approach, maybe.]
JWB> At some point, the SpamAssassin/quarantine model breaks
JWB> down...when you quarantine 10 large messages to let one
JWB> smaller message go through, you're somewhere close to that
JWB> point. As it is, we're busily installing four machines to do
JWB> the work that one would do quite well in the absence of
JWB> spammers (and they'll have help from another machine or two
JWB> so that users can see their quarantined mail and rescue their
JWB> false positives). And yes, SpamAssassin is part of that
JWB> picture.Some of the work the python.org postmaster (Greg Ward) is doing involves examining and rejecting messages during the SMTP dialog. The trick is to force them to do more work to deliver their spam, upping the costs of the sending host just enough for it to be not worth it. Unfortunately, I suspect our own costs will go up faster than theirs so we'll likely lose this battle too. But IMO, anything that isn't economics based will fail.
JWB> Plus, I fear the "new breed" spammers (the ones who actually
JWB> think their advertising is useful and welcome and only sent
JWB> to opt-in lists, although they buy the lists from guys who
JWB> [figuratively] sell them from under a trenchcoat at the
JWB> entrance to a dark alley) will cause legislation to be passed
JWB> forbidding filtering at the mail server level. It nearly
JWB> happened last time around.Yup, but on the other side of the coin, they're getting sued and losing for not verifying their opt-in lists. But none of that matters either given the global nature of things.
JWB> Ah, well...we'll see how it goes.can-i-be-a-rock-star-now?-ly y'rs, -Barry