[Mimedefang] Re: Justifying greylisting to management

[David F. Skoll]

Tomasz Ostrowski wrote:

> But it is also within its rights to issue 4xx. And if it makes it
> more reliable then why not?

Well, it's a matter of philosophy, I guess.  I can't see sendmail.org
accepting a patch for this.  Look at it this way:  Sendmail has every
reason to assume that if an SMTP client has a broken implementation of
the state machine on one message attempt, it probably will break on every
attempt, so why tempfail?

> This would not encourage the developers of broken servers to fix them
> or administrators to migrate. But it could be better than tempfailing
> after "data" because tempfailing "rcpt to" sometimes does not work -
> it will not waste bandwidth.

True; if bandwidth is a scarce resource, this could be an issue.  It isn't
for us, and I suspect it isn't for most people -- I doubt e-mail uses
the majority of bandwidth at most organizations.  Web browsing, P2P
and FTP probably swamp e-mail.

Regards,

David.