Charles Lindsey <chl(_at_)clw(_dot_)cs(_dot_)man(_dot_)ac(_dot_)uk> writes:
(1) A specification published in an accessible place. That might be a
web site, an Internet Draft, or even an experimental RFC.
An Internet Draft is clearly not sufficient to register a header for
standardization because it expires and the registry entry doesn't.
We could even go further and say that it was the ONLY route (that seems
to be the position Russ is taking)
No, it's not.
But if principles (1) and (2) were built into the system, many of
Keith's objections would be satisfied. People would know what the
current specification was. If they thought it was a dumb idea, or needed
improvement, there is a place to make their concerns heard. And, because
it is clearly provisional, they have ample warning that the details may
change, or even that the idea might be dropped altogether (BTW, expiry
of a draft without a durther draft should be automatic grounds for
removal from the Provisional Registry).
I don't see any point in putting details that may change or be dropped
entirely into an IETF header registry. If that's what you want, just use
Dan's registry.
--
Russ Allbery (rra(_at_)stanford(_dot_)edu)
<http://www.eyrie.org/~eagle/>