Kyle Hasselbacher <[EMAIL PROTECTED]> writes:

> I think SPF is great.  I think if more sites heeded SPF records I
> wouldn't get more bogus bounces than spam, but I'm just guessing.

I wasn't referring to SPF.  Sender address verification is something
different.  See for example

http://www.porcupine.org/postfix-mirror/newdoc/ADDRESS_VERIFICATION_README.html

An implementation for TMDA is included in the the distribution
(contrib/smtp-check-sender).

> I think also that the situation that SMTP-based C/R works "better" for is
> this:
> 
> * Spammer forges from a working address.
> * Spammer connects directly to the victim's mail server.
> 
> When both of those are true, TMDA sends a challenge to someone who didn't
> ask for it, but SMTP-based C/R doesn't.

Agreed.  The question is whether these fringe cases are worth the
additional complexity and difficulty of integration and installation,
not to mention the unforseen problems and gotchas.

> As you note, sender verification can solve the unwanted challenge
> problem for TMDA (and, incidentally, for SMTP C/R also).

Yup.

> If the spammer forges from a broken address, the difference between
> the methods is that TMDA makes it look as if the message was
> delivered.  With SMTP C/R, the spammer sees a rejection.

This assumes the spammer sits around looking at all these rejections.
I'm not convinved this is the case.

> Yeah, except TMDA's challenges are easier to read and to answer.

Which is pretty darned important.
_____________________________________________
tmda-users mailing list ([EMAIL PROTECTED])
http://tmda.net/lists/listinfo/tmda-users

Reply via email to