Re: [dmarc-ietf] Ticket #11 (and #112) - Proposed language

["Murray S. Kucherawy" <[e-mail]>]

On Thu, Jun 17, 2021 at 9:18 AM Douglas Foster <
[e-mail]> wrote:

> This is frustrating.   NP is a new design and the design issues should
> have been discussed before we got to this point.   I don't know why so many
> people are eager to not define the new technology.
>

Please don't conflate "I still don't understand what you're talking about"
with "I am eager not to define NP" or "I am unwilling to acknowledge
(something)".  They're not the same thing, and only the former is true.

Your message continues to assert something abstract.  I'll repeat my
previous request:

Could you craft a sample message (including DKIM details), sample envelope,
and sample DNS context (including A, AAAA, MX, and SPF details) that
highlights the problem you're talking about?  Maybe then it'll snap into
focus.

NP is an effort to partition the RFC 7489 SP=NONE result set, so that a
> subset of all SP results can be blocked on some additional criteria.
> This additional criterion could be based on non-existent as indicated by
> NXDOMAIN, or it could be based on "not used for email" based on a criterion
> to be defined.   Either approach needs to have a mechanism for
> non-compliant names to be made compliant.   I believe that this should
> involve a DNS entry, but the compliance measure should be specific to the
> NP test.   Requiring that every FROM address be linked to an IP address
> does not meet that requirement.
>

If I squint at this, maybe I can see what you're trying to argue: "
example.com" can't publish an "np" policy if they use a subdomain of "
example.com" that has no representation of any kind in the DNS.  Is that
correct?  If so, do you find this to be a defect, or simply a limitation to
be documented?  If you think it's a defect, why is that so?

-MSK