--On Friday, September 11, 2015 16:11 -0400 Scott Kitterman
<scott(_at_)kitterman(_dot_)com> wrote:
On Friday, September 11, 2015 03:34:26 PM John C Klensin wrote:
...
* And, while it is separate from the above, describe
the experiment to be performed, how it will be
evaluated, and any issues that might arise in
performing the experiment in a "live" Internet
environment (including any measures needed to back
away from it if it is not successful).
...
To pull out this one point...
I tried to discuss this in the WG, but didn't get very far
[1]. I agree the experiment in the experimental draft is
underspecified.
Scott K
[1]
https://mailarchive.ietf.org/arch/msg/dane/4gMgt2MiYLWYTmP-mOg
cxuOqqxg
While I continue to believe that there are substantive issues
with this proposal and that the document would need work even if
those substantive issues were dismissed or overcome, it seems to
me that there may also be a procedural one in that there have
been several claims about attempts to raise issues in the WG
that, to quote the above, "didn't get very far". I just looked
at the shepherd's report
(https://datatracker.ietf.org/doc/draft-ietf-dane-openpgpkey/shepherdwriteup/).
It, interestingly, still indicates that the WG wants Proposed
Standard, and that "The working group consensus is strong about
advancing this document.". It does not mention what one can
infer from the IETF was lively, and still unresolved, discussion
about a number of issues, including requested status.
It also indicates that this particular hashing mechanism was
adopted because the email community was asked for reviews and
that more review from them would be welcome. I don't know what
was tried; I do know that some of the comments to the effect of
"tried and didn't get very far" came from people who are
considered members of the "email community" and that there were
no requests for specific comments addressed to, e.g., the
standing SMTP mailing list.
The shepherd's report also asks for 'More review from email
community will not hurt, but unless they have an sudden insight
as how to "cannonize" email address this is the best we can do'.
As has been noted in several comments during Last Call,
canonicalization of email addresses (much less efforts to
'cannonize' them) is prohibited by RFC 5321 (and 821 and 2821),
so this is not going to be a matter of insight, at least until
or unless the IETF produces a standards-track protocol that
updates and replaces those SMTP provisions with specific
required local-part formats and rules. The comments in the
shepherd's report imply that the WG was not aware of those
restrictions. Independent of the substance of the matter, that
is a serious procedural problem. If people tried to raise it in
the WG and were ignored or dismissed (even in the shepherd's
report) that is even more serious.
I believe that, even if there were no other problems, those
procedural issues should cause the IESG to return the document
to the WG, request the document more consistently reflect its
proposed status, and get a shepherd's report that more
comprehensively reflects issues that have been raised (either in
the WG or in LC) and how they were resolved, rather than
claiming "strong consensus" about issues that apparently remain
controversial or expecting the community to come up with a
canonicalization approach without commenting on the relationship
of such an approach to SMTP requirements.
john