My thoughts about the problems of the IETF
moore at cs.utk.edu
Thu Apr 17 18:15:25 CEST 2003
> So, Keith's explanation pre-supposes that an AD will vote "yes"
> on a protocol action s/he has brought to the IESG. Is that
> in conflict with Scott's assertion that an AD overriding a
> WG's consensus about a document is a violation of process?
The thing to realize about the current process is that the rest of the
IESG won't even look at the document unless the sponsoring AD has voted
"yes" -- even if the sponsoring AD doesn't think his opinion alone
should be enough to block the document. So if there's a disagreement
between the reviewing AD and the working group things can get stuck at a
point when there's no visibility, no timers, and the IESG as a whole
isn't in the loop to break the logjam.
Sometimes you can get the other AD in the area to vote "yes" so you can
raise the objections you find in your review.
On one or two occasions I voted "yes" for the record but included my
objections in the writeup. Sometimes this was sufficient to get other
IESG members to review the document also, so that they could either
argue with my objections or agree with them (or add their own). On at
least one other occasion I got pushback on this and was told to work out
my disagreement with the(recalcitrant) working group - the IESG didn't
want to deal with it.
> Suppose the AD is the only IESG member who objects to a
> protocol action? Wouldn't the requirement that an AD
> support a protocol action amount to giving the AD unilateral
> veto power?
I don't think that's the intent - the intent is to make sure that things
that are on IESG's table have had preliminary review for the
requirements in 2026. (even so when I was on IESG, far too many
documents made it to IESG without having met basic requirements such as
a security considerations section, and I was as guilty as anyone about
failing to check for those.)
This glitch is an unintended consequence. Another unintended consequence
is that the document can be subjected to have multiple review/revision
cycles after leaving the working group over even trivial matters, even
when the working group is cooperative - for instance, if the reviewing
AD catches some things and the rest of IESG catches others.
More information about the Problem-statement