Hallam-Baker, Phillip wrote:
It seems to me that there are two possible effects a limited
charter can produce. One that precludes further work in a
given direction, and one that allows that further work to
happen in the future. Now, nobody's omniscient so the latter
is always a calculated risk, but it sure seems to me that
DKIM with respect to reputation is in the latter rather than
former category. I for one would be rather dismayed if it did
not lend itself well to reputation/accredation systems. It
seems to me that the risk is pretty low, so divide and
conquer seems more prudent.
There is a third effect that I have seen in WS-Security. The original
charter is so over-constrained leading to excessive debate over minutiae
and the whole process is slowed rather than speeded up.
I guess I'm less fearful of that than opening an entirely
new and decidedly less investigated area as part of the
wg's deliverables. I think that with ssp and auth-res on
the plate too we have quite a lot to get done. We also have a
requirements draft. Maybe that's the right place to put
hooks to make dkim rep/accredation friendly? Frankly I'm
not sure I know what is currently missing in DKIM to build
a name rather than IP based reputation system on top of
the DKIM foundation.
Mike