"John" == John C Klensin <john-ietf(_at_)jck(_dot_)com> writes:
John> On the other hand, if there is a real commitment to action,
John> then WGs have to be accountable for design decisions that do
John> (or do not) support the goal and be ready to explain their
John> decisions, even privacy-protecting ones that impose or
John> increase costs to performance, operations, or elsewhere. And
John> I would expect (not merely fear) ADs to push back strongly on
John> a WG that was unwilling or unable to do that and expect
John> Nomcoms to hold ADs accountable if they did not enforce the
John> intent of the rules.
Strongly agreed.
And if we don't have a community commitment to do that can we please be
honest with the world and tell everyone that when we look at the cost of
this issue it was something our community did not choose to pay?
John> It is not clear to me that the community is really willing to
John> make that commitment. Fortunately, I'm lucky enough to not
John> have to make that call.
I hope we are willing to make that commitment.
Note however that I do not want ADs pushing back at the end of the
process (unless WGs clearly failed to get adequate review of their
architectures up front.)
End-of-process security work is rarely all that productive.
You spend weeks or months fighting for little gain and for everyone
involved walking away with a sence of frustration and a diminished sense
of joy in the engineering challenges we face.
I want ADs to be pushing back in the early architecture phases.
Comments on early architecture drafts, and WG-decision-level appeals
filed early in the process are probably better tools for pushing back
on WGs that are not adequately considering privacy than late-stage
discuss positions requiring architectural change.
And as Eliot points out, the question of what balance in tradeoffs is
appropriate will evolve over time. At the beginning if a WG does a good
job of considering something and you just don't like how they balanced
the issues, your only option is to start a broader discussion. That's
how it should it be. If community norms emerge, then there are tools
for applying those.