[dmarc-ietf] For the chairs: The lack of consensus

[Douglas Foster <[e-mail]>]

It seems clear that we have lost consensus.  The next DMARC-related RFC
will not be able to justify standards track unless a miracle occurs, and a
miracle seems unlikely.

The problem was evident years ago, when Dave Crocker’s proposal was on the
table.    It captured all of the frustration of the mailing list
community.   He was convinced to go quiet, but the mailing list problem was
not solved.   This previous milestone promised ARC as a solution, but that
promise has not been fulfilled.   So we have a potentially large body of
opposition at the next milestone.

Internally, we have also had a lot of opposition.   Too much of this
discussion has been combative rather than cooperative.   As one painful
example, it should not have required a year of complaining for me to obtain
an open discussion about the definition of non-existent domains.
Additionally, I know that we have lost some participants because of the
tone of the group.    I also note that we have lost a lot of active
participants in total, for reasons unknown.   Those of us who remain are so
few that it is bold to presume that we represent the interests of the
legitimate portion of the email community.

Then, Michael Hammer warned us that DMARC had a security vulnerability when
sibling authentication was used within unidentified private registries.
We should have immediately moved to deprecate sibling authentication.
Initially, we dismissed the problem as too difficult to solve, then pursued
the tree walk algorithm which exacerbates the problem.

And now, the tree walk draft has been developed by a process which has
openly discarded collaboration.  It should be reclassified as an individual
submission, not a working group document.  Several of its assumptions are
disputable, but disputing has been unproductive.   The last week of silence
indicates that its advocates have no intention of changing course.

At present, my agenda is to develop a specification which integrates the
PSL with domain owner indicators and evaluator knowledge, without
discarding any data.     I hope that the effort will be assisted by some
members, but I have no hope of acceptance from those who are already
committed to other outcomes.

Consequently, I am asking the chairs to allow project forking to create
divergent individual (and subgroup) drafts.   Hopefully, the number of
unique proposals will be small.   If consensus remains as divided as it is
now, those can move forward as individual informational documents looking
for enough interest to become experiments.



 Doug