I agree with this analysis. Some additional comments below...
--On Wednesday, 16 November, 2005 10:04 -0800 Lakshminath
Dondeti <ldondeti(_at_)qualcomm(_dot_)com> wrote:
At 08:29 AM 11/16/2005, Jari Arkko wrote:
I am not convinced that we need to do any change in this
area. First, I think we have higher priority items to worry
about. The IETF really needs to spend its process change
cycles on things that provide a measureable effect and that
has a real impact on timeliness/quality/openess/your favorite
As to whether this is urgent is a good question. But, the
change seems straightforward enough that we can accommodate it
as part of "regular business."
It is useful to do it now for one additional reason and that is
precisely because, at least to the best of my knowledge, no
recall actions are pending or in process. We don't want to be
rethinking these sorts of things at a time when doing so raises
"who are we trying to get" questions.
And we've never had a recall so optimizing process
for it is premature. Lets wait until we have recalls
and then tune the parameters.
Well, if this argument is extended backward, we wouldn't have
the "20 signature" rule at all, but a "1 signature, signed by
'anyone'" rule. Perhaps that would be better, perhaps worse. I
have been thinking that "one" is easier to get than ten or 20,
but I find Harald's "safety and sanity in numbers" reasoning
fairly persuasive. But the point is that we have already done
some tuning/ process optimization in the absence of experience;
adjusting that process on the same basis seems plausible.
I do believe that if the IESG or IAB has a problem working
together then we should hear about it as soon as possible.
But nothing prevents open discussion, including participation
of IESG members. If the only people who can be convinced
to sign a recall petition can be found in the IESG and IAB,
well, maybe we don't have a problem after all.
Think about it this way. Recalls are a nasty, horrible, very
burdensome, and disruptive mechanism. That is true whether the
petition threshold is one signature or two dozen. At the same
time, they are the only mechanism we have. As has been pointed
out at least once on-list, indirectly by Lakshminath below, and
several times in private notes, one of the values of a recall
effort that starts is that it lays the foundation for a
completely unambiguous "this is serious not just random whining;
you need to either clean up your act, resign, or this petition
goes to the ISOC President and all the dirty laundry gets hung
out in public" conversation. That isn't attractive either, but
it is much better for the community than an actual recall. And,
for that purpose, a petition with several signatures, with
people from both within and outside the relevant body, is far
more persuasive about the reality of serious discontent than one
Finally, I'm not sure I am comfortable with the idea that
the IESG and IAB (altogether over 20 persons) could
all by itself sign a petion and get it processed. This would
appear to provide a channel to oppose decisions from the
nomcom, for instance.
The confirmation process takes care of this already. The IAB
doesn't have to accept the slate of IESG nominations the
nomcom sends to them, and so on.
But the IESG and IAB don't get to approve nomcom selections to
the IESG and IAB, respectively. But think about this: If the
IESG and IAB feel strongly enough that someone is a bad actor
that they can get majorities of both bodies, or most of one and
a few from the other, to sign a petition to initiate a very
nasty process that would be almost certain to result in a
low-performance slot while the process plays itself out, maybe
the odds that the nomcom has made a mistake are high enough that
the community deserves the opportunity to do some
Remember how the appeals process works: the ISOC President needs
to get a petition. The secretariat needs to confirm the
signatures, which probably doesn't happen overnight. The ISOC
President appoints a recall committee chair. We've never been
through that process, but I'd assume, based on experience with
Nomcom chairs, that the ISOC President would do some
consultation and that the first person selected might say "no".
Figure that, by the time we get signatures verified and a chair
appointed, it is a couple of weeks after the petition
submission, and that is pretty optimistic. Then, just like the
nomcom, the Chair asks for volunteers from the nomcom-eligible
pool, presumably less the sitting nomcom members (RFC 3777 is
not clear about that, and should probably be fixed). The
selection process then requires a 30 day volunteer period,
followed by a one-week waiting period, followed by selection,
followed by a one-week challenge period. Note that we are now
at least eight weeks post-petition and the recall committee
hasn't met yet, much less done anything. But, unless the "one
month to get organized" of RFC 3777, Section 4.18, is applied to
the recall committee, it is at least ready to start work.
Now assume that the recall committee can do its job in a couple
of weeks (which I would consider a miracle) and concludes that
the person should be removed. Circa ten weeks out, the Nomcom
Chair convenes the Nomcom (most of whose members may believe
that they were finished for the year) and starts the process of
soliciting candidate nominations, evaluating them, interviewing
would-be candidates and listening to the community, etc. Once
they have a nominee, the nomination has to go through the
conformation process. My guess is that it would be a minimum of
five or six _months_ between initiation of a petition and the
time we have a replacement person in place. During the first
half (or so) of that time, we've got someone in place who is
probably mad at everyone and possibly continuing to do whatever
got him or her in trouble in the first place only more so.
During the second half, we have a vacancy.
I suggest that anyone who is willing to initiate a process like
that would need to be quite sure that things are serious enough
to justify it, either in terms of serious bad behavior or in
terms of needing to establish a principle. With 20 (or even
ten) signatures required, there need to be a lot of such people.
Within the IESG or IAB, where there is a keen understanding of
the implications of operating short-handed for many months, the
threshold is even higher. And, as Lakshminath points out,
anyone signing these things puts their own credibility and
quality of judgment out for evaluation by the community, for
better or worse.
The IESG and IAB have something to lose, if you think about
it, if they go about launching spurious recall petitions.
Their names will be published along with the petition and if
the recall committee comes back saying there is no need to
recall, then people will take a hard look at the petitioners.
That's a good deterrent, I think.
Agreed. See above.
Ietf mailing list