Hi Keith,
Perhaps the proponents for the WG looked at what IDR was doing, thought that seems to work well, and they should try something similar here? I.e., there does seems to be some supporting analysis that this approach may work well for mature, widely deployed
technologies. Is there any recent evidence that this approach has been tried and hasn’t worked?
But equally, maybe this feedback would have been better during the community review phase in April, i.e., BEFORE the WG was chartered? I.e., the premise here seems to be that this
is such a terrible precedent that the IESG should ignore the normal rules and process for chartering the WG and make up some new ones on the fly that allow the charter to be modified after the WG has completed all the steps specified in the process. This
hardly feels justified in the case where we don’t actually know that the current charter is a problem, only a supposition that it will be.
I can’t see why the best option here isn’t just to try it out and see what happens. If the requirement turns out to be too restrictive then the WG can always be rechartered to tweak
the process or remove it altogether. But I’m not even convinced that there will even be an issue here at all ...
Regards,
Rob
From: Keith Moore <moore@xxxxxxxxxxxxxxxxxxxx>
Date: Friday, 24 May 2024 at 14:41
To: ietf@xxxxxxxx <ietf@xxxxxxxx>
Subject: Re: [Ietf-dkim] WG Action: Formed Mail Maintenance (mailmaint) / Commitment
On 5/23/24 15:47, John Scudder wrote:
> Sometimes we should be brave enough to try new things. Maybe this is
> one of those times?
I don't think bravery is the issue here. The question in my mind is,
does this actually help make IETF standards developed by this WG more
relevant to the Internet?
So far, I fail to see how it does.
Existence of independent prototype implementations is a nearly
necessary, but far from sufficient, condition, for ensuring that.
(I say "nearly" necessary because it's actually possible for a protocol
to be successful in market terms even if there's really only one
implementation - heck this is almost true for web browsers today.)
Or to put it in terms of bravery, sometimes we should be brave enough to
NOT do whatever strikes someone as a good idea, without first doing some
analysis to see whether it actually helps. And I mean no criticism of
whoever thought this was a good idea. But fundamentally the practice of
engineering is to use analysis to determine what will work well, BEFORE
it's actually implemented.
Keith
