It's fairly easy to demonstrate interoperability of protocols, but
usefulness is much more difficult. DKIM is an infrastructure protocol,
designed to provide a basis for other mechanisms, such as domain-based
reputation, to operate. Those other mechanisms are as yet nascent; how
does one judge usefulness at this point?
This appears to be imposing criteria that go considerably beyond the
IETF's requirements for Draft.
So is it your position that a protocol must be advanced if it meets a
minimal set of criteria even in the face of engineering judgement that the
protocol is not yet sufficiently deployed to have a reliable understanding
of the adequacy of the current design?
As I see it, the reasons to go to DS would be
Y1. to progress a fairly stable standard along a defined track, and
Y2. to review it and perhaps clean it up a little along the way, and
Y3. to get broader deployment as a result of higher maturity.
As to Y3, there's evidence, as Murray has pointed out and as many of
the rest of us are aware, that most deployment comes from publication
as PS, and from other sorts of publicity... and DS probably doesn't
create the swell of deployment that we might like. Still, as long as
the IETF considers the three-stage standards track to have value, I
think there's some value in working within it.
The reasons not to go to DS would be
N1. to avoid wasting our time on nominal advancement that has little
or no real value, or
N2. to avoid wasting any more time working on something that's not
very useful, or
N3. in recognition that it's not stable, and that, while it certainly
meets stated criteria for DS now, we think we're likely to change it
significantly after more experience with it.
My opinion is that N1 is arguable, but that N2 and N3 are not the
case, and that we shouldn't resist advancing DKIM base to DS for
reasons N2 and N3. My opinion is also that, while N1 might be true,
the fact the IETF considers it worthwhile overrides that.
And note that I'm only advocating advancement for DKIM base at this
point; I think we DO need more experience with ADSP before we have any
clue whether it's stable (or useful).
Barry, as participant
NOTE WELL: This list operates according to