At 12:55 PM -0800 11/23/10, Doug Barton wrote:
In the theoretical perfect world the reason for producing a spec is so that
vendors can _create_ interoperable versions of the service. That motivation
doesn't apply here, so one wonders what the time pressure is.
That is "a" reason, not "the" reason. Another is so that vendors can assure
interoperability of their systems after they are deployed. Saying "you have to
figure out how to interoperate with all other implementations" does not lead to
that interoperability. A stable RFC can help greatly, as has been shown
repeatedly in the IETF.
--Paul Hoffman, Director
Ietf mailing list