On Feb 7, 2017, at 1:54 PM, John C Klensin <john-ietf(_at_)jck(_dot_)com>
wrote:
TL;DR summary: Just say 'no'.
No hats, personal opinion only, and in response to several of the messages in
this thread, not directly to John.
Will it ever be possible for new work to be proposed at the IETF without
drowning out a chorus of nitpickers? You may not want to help standardize it.
You may not want to implement it. You may not want to deploy it. You may not
want to use it.
So don't.
This work won't hurt the Internet, particularly if done with productive input
from a couple of people with experience in the existing system, and some
cross-area review.
You may not like that all application developers want to use HTTPS (and
derivatives like QUIC), WebSockets, JSON, etc. However, forcing everyone who
wants to standardize a modern protocol to justify their desire to use a set of
modern building blocks has gotten tiring, and makes people avoid bringing good
technical work to the IETF that would benefit from cross-area review.
Every time we make this sort of work difficult at the IETF, we reinforce the
message that the IETF only exists to pay homage to the protocols of our
forefathers. When it comes to email in particular, the system that accreted
over the years is baroque at best. We should not stand in the way of any
attempt to simplify the steaming pile of cruft we preside over because the work
might be hard.
Several of the concerns that have been raised in this thread are interesting,
useful, and might make the work better over time. Had they not been phrased
with an implied "so therefore your whole idea is dumb and you should go away",
I probably would have been able to keep myself from writing this note.
TL;DR summary: Just say 'yes'.
—
Joe Hildebrand