spf-discuss
[Top] [All Lists]

Re: Re: For SPF council review: Syntax error = Perm error = Message should be rejected?

2005-05-02 19:19:00
...... Original Message .......
On Tue, 03 May 2005 00:06:53 +0200 Frank Ellermann 
<nobody(_at_)xyzzy(_dot_)claranet(_dot_)de> wrote:
Scott Kitterman wrote:

I feel pretty strongly that rejecting messages from a domain
with a malformed SPF record is a really bad idea

I think that it's an excellent idea.  Otherwise you get hard
to find intermittent problems, sometimes it apparently works,
sometimes it doesn't, and in the worst case mails are lost.

It's one of the core ideas of Wayne's drafts to catch errors
a.s.a.p. and to report them as clear as possible.  That was
the very reason why Wayne risked to split and kill the SPF
project with his "unofficial draft" in parallel to Mark's
official emergency draft after the demise of MARID.

But he convinced us, or at least me.  You read spf-help, new
users have fantastic ideas how to create erroneous policies.

Add implementations desperately trying to "be liberal in what
you accept", and the result is an unpredictable chaos of half-
working policies,

People new to SPF make all kinds of mistakes.  If after the
first mistake, they start getting messages rejected, they'll
just give up and go home.

They don't get a FAIL, they get a PermError "fix your policy".

If their polic is broken but apparently works for some cases,
or a tolerant implementation, or a receiver not checking SPF,
but later it sometimes doesn't work, then that would be worse.

Just think about a syntax error tolerant compiler, or the pain
of Web pages apparently working only with some browsers.  It's
IMHO a nightmare.  Wayne's concept "validating implementation"
is hard, but clear, and reliability is very important for SPF.

                       Bye, Frank

I'm OK with whatever the council decides.  I do think that rejecting mail 
from domains with problematic records is draconian and there are softer 
methods available.

Scott Kitterman


<Prev in Thread] Current Thread [Next in Thread>