Ned Freed <ned(_dot_)freed(_at_)mrochek(_dot_)com> writes:
Like it or not, careful reviews and review checklists, while quited
flawed in their own right, are the best tool we have. When I was on
the IESG I had my own private review checklist; it was the only
thing I found that worked.
I agree careful reviews are necessary. What I find surprising is your
logic, which seems to say:
IANA considerations sections in IDs are not sufficient, therefore
they are useless (or worse).
I have never said or even implied that.
Is that really what you are advocating?
Of course not.
What exactly is it that you
think should be done (in addition to careful reviews) that would help
reduce the odds that the careful review find issues with the IANA
instructions (or lack thereof)?
Simple: The requirement that an IANA considerations section be included in RFC
not containing any IANA actions needs to be dropped. I have been extremely
clear ahout this.
Note that having IC sections is all about improving the odds that they
contain the Right Thing before the document is approved by the
IESG. In my mind that means:
That may be the intent, but the effect is to substitute boilerplate for
review, which won't improve specification quality at all.
1) IANA reviews an (essentially) final version, to be sure what it
says is consistent with their understanding of what needs to be
carried out. But, IANA does this review during Last Call. Thus, the
IC section really needs to be complete _before_ the full IETF
2) Well, the Shepherding AD can do the "careful review" during the AD
review phase, but there is already plenty of pressure to skimp on
the AD review in order to send a document the WG says is finished
to IETF LC ASAP. I.e., to get the IETF LC started and "fix any
issues that come up later". Plus, in my experience, plenty of IC
issues are caught by ADs other than the shepherding AD. So relying
on them to catch all such issues is far from ideal.
3) Voila, have a checklist item that alerts WGs to things they should
take steps to make sure their documents have already addressed
prior to advancing a document out of the WG.
This is all very logical, but we're dealing with people here, not perfect
Ietf mailing list