On 6/28/23 8:40 PM, Christian Huitema wrote:
On 6/28/2023 7:02 PM, Keith Moore wrote:
Strong consensus shouldn't be a reason to relent. I mean, why have
the IESG review it at all if a discuss can be overridden by "strong
consensus". When I brought this up in the wg way back then (ie,
~2013), it sort of reminded me of an echo chamber which happens to
some wg's from time to time -- it wasn't just the main author
raining fire on me. I've always thought of the IESG as sort of a
backstop for when that happens along with its function of tightening
things up. I guess it also begs the question of why it was allowed
into their charter in the first place.
It absolutely is the IESG's job to push back on documents of poor
technical quality even when there is "strong consensus". Or to put it
differently, the criteria for Proposed Standard are not merely "rough
consensus" but also "no known technical omissions". To qualify for
Proposed Standard status, both criteria need to be met.
Keith, you may remember that one of the big gripes against the old IAB
is that it sometimes acted as "technical dictator", and would try
impose specific technical decisions against the will of the working
group, or of the IETF in general. Do you want a repeat? Having a
managing body push back against a "strong consensus" would probably
not end well.
BCP's are not the same creatures as standards that involve bits on the
wire. In this particular case it is a BCP saying that bad guys should be
good. Two IESG members saw that as it turns out, yet it was published
anyway. That is an embarrassment not only for the working group that
produced it, but for the IETF as a whole. A standard that doesn't
achieve interoperability or doesn't achieve its own goals gets what it
deserves and the fault lays squarely on the working group that produced
it -- it's not the IESG's job to micromanage working groups. But
something like this is a reflection on the whole IETF that we are
complete idiots. What the IESG should have done in allowing it to be
published at all would be to publish it on April-1 for some plausible
deniability.
Mike
