Joel,
While others have touched on pieces of the response I'm about to
give, I feel some obligation to respond directly to my
understanding of your point.
--On Sunday, October 20, 2013 15:54 -0400 "Joel M. Halpern"
<jmh@xxxxxxxxxxxxxxx> wrote:
> The obvious risk if one separates WG management from review is
> that one could easily get the situation where the manager, in
> working with the WG, says that the document needs X, Y, and
> not Z. And then the reviewer says "needs Z". And there are
> more extreme versions of this.
Noting Dave Crocker's concern about the use of terms like
"manager", very little would change that would affect your
scenario. The decision as to when a document was ready to go
into IETF LC would remain up to the relevant AD and IESG
(however the IESG decided to handle that which, as you know, has
varied over the years). So, if the relevant AD said "needs X,
Y, and not Z", he or she could presumably keep the document in
the WG until either that requirement were met, he or she was
talked out of it, or the issue was appealed. ADs would continue
to be fully aware of and to a considerable degree, accountable
for, "their" WGs and the quality of their technical work
(presumably including reviewing shepherd's reports), so that
part of the workload doesn't change.
If the IESG decided to force full IESG review and approval prior
to IETF Last Call, this family of proposals would not reduce
their workload at all and might ultimately increase it. At the
other extreme (and preferably), if the relevant AD reviewed
materials to be sure everything was in order (procedurally and
technically), consulted other ADs, directorates, etc., as she
felt needed to be assured that things were in order, worked with
the WG and its leadership on what appeared to be outstanding or
problematic issues, and then forwarded the specification for
IETF Last Call, we would eliminate both the intensive,
word-by-word IESG reviews of documents, the need for the IESG to
take a "vote", and the requirement that ADs either perform or
arrange for extensive cross-area reviews (or carefully review
out-of-area documents themselves). Note that also implies that
the IESG wouldn't get tied in knots over DISCUSS positions on
document approval: it would just not be their job and, if ADs,
especially ADs outside the relevant area, had concerns, their
obligations would stop when they brought those concerns to the
attention of the review body (with AD concerns presumably
carrying more weight than random comments on the IETF List
during LC, but not much more than a community LC comment that
was equally well justified).
And the responsible AD would be in a position to explain or
advocate for the WG's work/proposal to the review panel or could
delegate that job completely to a shepherd with a
slightly-revised "job" description.
Much of the management (sic) of IETF Last Call would fall on the
review body (details to be worked out). That body could, as
others have suggested, solicit additional reviews from
individuals or groups, including asking ADs for advice on the
right people or groups to do that. But they would bear the
final review and approval responsibility and would be directly
responsible to the community for it. If, using your scenario,
they said "do Z", it would be no better and no worse than a
responsible AD imposing requirements X and Y before a Last Call
was conducted and then the IESG coming along and imposing
requirement Z after IETF LC completed -- modulo quibbling about
the difference between "requirements" and "comments to which
you'd better pay attention", that happens, a lot, today.
> Currently, the ability for the managing AD to say "no, this
> won't pass muster" is part of his management tool. If he is
> not the reviewer, he seems to have lost an important tool. I
> hope I am missing something that would make this sort of
> approach workable.
Again, the responsible AD still has to clear a specification for
IETF LC, so that doesn't change (regardless of what one thinks
about "management").
best,
john