John C Klensin wrote:
Two observations, one substantive and one procedural...
Substantive: Yes. Regardless of how one feels about this behavior in
an alternate (perhaps even better) universe, in this one, there is just
too much established practice consistent with Ned's description for
second-guessing it to work. If someone wants something else, it is
probably time to dust off and finish some flavor of per-recipient
post-DATA reply model and see if it gets any traction.
You might as well revert your changes in 2821bis back to the mistake
as it was written in 2821 - 5zy does allow client retries. It is
clear there are software trained around that 8 year old terrible mistake.
Of course, no one is going to agree anything is wrong and there is no
way anyone that will even dare to go against the Neds to join me when
I right here.
Very sad.
--
Sincerely
Hector Santos, CTO
http://www.santronics.com
http://santronics.blogspot.com