Arvel Hathcock wrote:
The phrasing of the query for a poll is a challenge, here.
Nope. The question is on the removal of the NXDOMAIN step from the
working group draft.
That's one question, yes, but there are more. While I'm not sure it captures
the available choices, here's an attempt:
There really are two sets of choices:
Degree of requirement
---------------------
1. Make a domain 'validity' test mandatory.
2. Make a 'validity' test optional (SHOULD vs. MAY)
3. Remove a 'validity' test
Type of 'validity' test
-----------------------
1. NXDomain
2. RFC2821bis
I suspect there are more choices for type of test, but these are the two that
are obvious. Unfortunately each of the listed choices has some problems. In
addition there is the problem that the industry already conducts these tests,
but in a variety of ways.
Basically, the industry does not have a solid, de facto NXDomain standard. It
has a small number of common choices, but 'common choices' is different from
'definitive, single choice'. If this working group insists on defining a
single, definitive test, then it is running some risks with industry adoption,
since we are attempting to constrain existing practice.
The RFC2821bis test potentially has multiple queries, meaning higher overhead.
But it is semantically deeper, with respect to validity for email-related use.
It also seems likely that the more definitively the working group attempts to
specify (standardize) this particular test, the more it will be subject to
review by the larger Internet email technical community, for architectural
impact. But, of course, that's just my own opinion.
d/
--
Dave Crocker
Brandenburg InternetWorking
bbiw.net
_______________________________________________
NOTE WELL: This list operates according to
http://mipassoc.org/dkim/ietf-list-rules.html