Barry is technically correct:  Any coding system can be technically
sufficient if it is explained.    We could, for example, choose

qzx=(#|&|~)  instead of psd=(y|n|u)

Of the infinite set of choices, how do we choose one, or does it even
matter?  The answer should be driven by the needs of the reader -- what
coding scheme will be most easily understood and used correctly by software
developers, and will be most easily understood and used correctly by domain
owners?   My opinion, No.

-1

This is symptomatic of the larger problem with our current draft.   Our
description of the DMARC evaluation process is very close to minimal
technical sufficiency.   By "minimal technical sufficiency", I mean that
there is nothing present or omitted from the document which prevents
creation of a correctly working implementation.  Therefore we could "ship
it", and perhaps we will.

But our goal should be to produce a document which simplifies the effort by
the reader.   This includes a developer's need to turn specification into
code, and a system administrator's need to turn code into better
disposition decisions.

To know that we have met the needs of developers, we would have to
be willing to evaluate multiple outlines and consider a rewrite.   To know
that we have met the needs of the system administrator, we would have to be
willing to rethink our scope.

Doug





On Thu, Jul 21, 2022 at 2:32 PM John Levine <[email protected]> wrote:

> Now that we've gone around the barn a few more times and nothing has
> changed with the psd=u/n/y tag, can we please ask for a tentative IANA
> registration?
>
> R's,
> John
>
> _______________________________________________
> dmarc mailing list
> [email protected]
> https://www.ietf.org/mailman/listinfo/dmarc
>
_______________________________________________
dmarc mailing list
[email protected]
https://www.ietf.org/mailman/listinfo/dmarc

Reply via email to