On 05/04/2011 08:16 AM, Dave CROCKER wrote:
Michael,
On 5/4/2011 7:58 AM, Michael Thomas wrote:
This is a good example of why this effort has come off the rails.
Going from 4871 to DS should have been a fairly straightforward
effort considering the high degree of interoperability we achieved.
Instead of just removing a few unused features, we've seen a
wholesale rewrite when one was manifestly not needed. Worse,
is that when that history is mentioned it is either disregarded
or sneered at by the senior editor. That is a problem.
"Wholesale rewrite"? Well, that should be easy for you to document,
given how convenient it is to point to the existing diffs. The task
when citing a problem with changes is to point to /specific/ changes
that are problematic. That requires some work.
In particular, please cite normative differences.
As I've already mentioned: the notion of "output" is a huge NORMATIVE
change. It should have never been allowed as a piece of errata, and
continues to cause all manner of issues. 4871 was NOT BROKEN in this
respect. We have no business telling developers what the layering
relationship is between the DKIM verifier and its consumers.
As for the particular difference between rfc4871's statement of DKIM's
purpose and rfc4871bis' statement, you might want to review the
language in the Service Overview and the Deployment documents. On
this issue, the working group's learning process has been incremental
and well documented.
Asleep at the switch is a more accurate portrayal. By the time any of this
effort started, we had a wealth of interoperability information which
resulted
in the non-controversial parts of the errata, and stable unchanging
implementations. The "learning" should be to leave well enough alone,
instead of changing things just for aesthetics and quite likely breaking
things in the process, either purposefully or inadvertently.
As for 'sneering' at history, please do cite that occurrence, too.
Please distinguish between citing history versus sneering at it.
Explaining the criteria that qualifies as 'sneering' would be helpful.
You need look only at the piece that I quoted from. Your continuous use of
ad hominem attacks is well documented.
Mike
_______________________________________________
NOTE WELL: This list operates according to
http://mipassoc.org/dkim/ietf-list-rules.html