Re: Spam wars (was test)

"bogus address" <[EMAIL PROTECTED]> wrote

> > I don't filter that way. I have been using SpamPal which uses blacklists
> > to do my filtering. The way I have it set, it tags anything suspect with
> > [Spam] and I use the OE filtering to move this stuff into its own folder
> > where I give it a quick review once in a while in case something genuine
> > has got there in error before doing a block delete.
>
> That isn't scalable.  My own setup is bad enough.  All my mail transfer
> agent looks at is the SMTP envelope (no filtering beyond "does the
> recipient exist?") and in particular I never even see the headers of
> suspected spams.  But over the last month the spam load has got to the
> point where most of the traffic between this machine and my ISP is used
> in rejecting spams; with this rate of increase, unless Demon can make
> the damn things disappear from their server without me ever seeing them
> at all (a service they refuse to implement) I'll have to abandon them
> as a mail provider sooner rather than later.  Your setup seems to be
> downloading either headers or complete messages before deciding whether
> or not they're worth reading, which means you're wasting 20-40 times
> the machine resources on each item of spam than I am and you're heading
> for meltdown.

Agreement, except that the tagging is done based on "known enemies" sort of
on the lines like "Loads of spam comes from this IP address, rather than
"This says 'hi' so it has to be spam". Given that 90% of spam seems to come
from known sources, I'd suggest it is the more logical approach.

Much as I hate the ideas of policing the Internet, I'd agree with a post
Chris Rockliffe (I think) made earlier in this thread. He didn't suggest
this but personally, I wouldn't mind seeing the professional spammers locked
up. (and I am generally against Internet policing)