On Wed, 14 Feb 2007 22:52:45 -0500 (EST) Dean wrote:
DA> As an exercise, try to show
DA> that Administrator A, acting as Story described, has acted unreasonably
DA> (as Lemon asserted)  when Administrator A can quote the text of this
DA> draft in his defense.

Let's do two exercises. Given a document with the wording your
propose, try and convince Administrator A that he MUST NOT use a system
that he's already determined to effective.

I think saying apps SHOULD NOT depend on reverse mapping is ok; the
rfc 2119 definition still allows administrators to decide to use it in
their environment, however unreasonable that seems to others.


DA> I do feel the wording is way too weak.  Basically, everyone should
DA> consider that if this document is approved as is, and Administrator A
DA> blocks your email because, e.g., it has the word "dialup" in the reverse
DA> mapping entry, then Administrator A will refer to this document as proof
DA> of the reasonableness of his actions.

They don't need this document for that. All they need is their own
statistics on the percentage of dialup or dynamic host spewing SPAM.
You keep going on about how reverse DNS isn't proof of security and
pointing your finger at me. But the example I gave is not using it to
provide security - it is using it as proof of insecurity. 

DA> Likewise, if the reverse mapping doesn't exist,
DA> or doesn't "match" the forward mapping, etc, they will
DA> also point to this document as justification for their behavior.

Forward Confirmed reverse DNS is an entirely different beast. I would
be in favor of the document have strong wording against this type of
check, because there are very good reasons why an ip address might be
associated with multiple domains. Anyone who tries using this method
would quickly for email would quickly find out it's a bad idea (see
http://news.com.com/2100-1023-982118.html).

-- 
Robert Story
SPARTA

Attachment: signature.asc
Description: PGP signature

_______________________________________________
DNSOP mailing list
DNSOP@ietf.org
https://www1.ietf.org/mailman/listinfo/dnsop

Reply via email to