Nomcom (was: Re: Requirement to go to meetings)

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

 



Subject changed, this is about to go off in a different
direction.

--On Thursday, October 27, 2011 08:38 -0500 Mary Barnes
<mary.ietf.barnes@xxxxxxxxx> wrote:

>...
> [MB] No, I do not think the comments should be public.  My
> point was that there is such a small percentage of the
> community that even provides input that it is really difficult
> to make really good decisions with that information.  Each
> nomcom asks for detailed feedback, but it is rare to get
> feedback that provides concrete examples of why person x is
> the best choice for a position.  That makes the job of the
> Nomcom extremely difficult and is one of the reasons why the
> decisions can be far from perfect.
> 
> The primary problem with the current Nomcom model that can't
> be fixed by a process change is the lack of community input
> (Note: there is still time for folks to provide that input for
> this year's Nomcom!!!)  So, I consider voting to be an easy
> way for folks to express an opinion without providing details
> - it at least could hopefully broaden community participation
> because it takes far less effort.

Mary,

Please understand that what I'm about to say is not a criticism
of any hard-working Nomcom or its members, present or past.

First, the problem with voting in our environment is ensuring
fairness: if nothing else, it would be really easy for a company
who felt like doing so to pack the proverbial ballot box.

But, more important, when the Nomcom model was developed, I
think the assumption of most of the community would be that the
Nomcom would be populated mostly by really active participants
in the IETF and that, in general, if a particular candidate
wasn't personally known to at least several Nomcom members, that
would be a really bad sign about that candidate.  In that
environment, extensive polls/ questionnaires were probably
unnecessary; at worst, Nomcom's didn't need to depend on them as
a primary source of information.

Since then, several things have happened.  The most important is
that the community has gotten bigger and a number of trends have
combined with size to make the assumption of first-person
knowledge a lot less likely.  The number of positions to be
filled has also increased significantly: for example, I'm pretty
sure there were fewer areas in 1996 and the typical area had one
AD, rather than two being the norm for everyone by the IETF
Chair.  The Nomcom process has gotten longer (probably in part
as a result) and serving on a Nomcom has become more burdensome.
With most positions to be filled by the Nomcom, effectively
having a rule that anyone interested in any of those positions
should not volunteer to serve on the Nomcom may have some effect
on participation.   I'm sure you could add to that list.

For that set of reasons, while I don't think voting is the
answer unless we decide to change our membership model so that
we can identify close affiliates and vote by organization and/or
restrict the ability of any organization or group of
organizations to capture the process --and, for the record, I
don't favor trying that -- I do think some fundamental
rethinking of the Nomcom model may be important and important
soon.

In the interim, if the Nomcom wants more feedback, I suggest
that some tuning (not requiring major reforms) may be in order
to actually make giving that input easier.  Let me give four
examples, but they are only examples.  There may be others that
would be even more useful to think about.

(1) There is an incumbent bias built into the system because
Nomcoms (using a reasonable reading of the criteria in various
BCPs), appear to believe that their job is to return an
incumbent who is willing to serve again unless a candidate is
available who would be clearly superior.  Especially for the
IESG, unless the incumbent has screwed up in a serious way, that
bar will normally be impossibly high because a sitting AD is a
known quantity and almost any alternate candidate is a risk.
Spencer Dawkins and I made a proposal a couple of years ago for
the Nomcom to review the incumbents who were willing to serve
again separately, make decisions about them, and only then start
considering others.  The proposal, at least in the last version
posted
(https://datatracker.ietf.org/doc/draft-klensin-nomcom-incumbents-first/)
never got an real traction.  The proposal had some
disadvantages, some of which might have been fixed with more
work, and others of which were just tradeoffs to be considered.
But one clear effect would have been to drastically reduce the
number of positions for which the community is asked to comment
on a list of candidates.

(2) Independent of that particular suggestion, the number of
slots on which people are now being asked to comment, and, in
some cases, the number of people who are listed for each slot,
is just daunting.  Ask someone to comment on two or three
candidates for one position (or even a few positions) and you
will probably get a lot of input, especially if the people you
ask seem to be a focused list, chosen in some intelligent way.
Ask folks to comment on 48 names and I suggest that human nature
predicts a lot of "maybe I'll get to that another time" and
general tuning out.  

Perhaps it means that RFC 5680 was the wrong answer to the
question/ problem, but I suggest that there is a reason why no
plausible search process in the work world presents a final
selection process with 48 people, or even seven, for in-depth
consideration and comments.  Developing short lists has the
advantage that the selection process gets comments on those
candidates for whom it really wants and needs comments _and_
people have more incentive to write good comments for the
smaller number of people for whom they will be useful.  Listing
everyone who expressed interest (or was willing to be listed for
some other reason) might get a Nomcom to look again at a
superstar who had slipped through the cracks of earlier
considerations but is much more likely to just overwhelm the
system with noise. 

(3) I don't believe it has happened this year, but trends in the
community predict that, sooner or later, we will have someone
volunteer for a large number of positions simply because he or
she (or whomever they work for) wants them in the IETF
leadership.   Someone ought to be able to write one comment that
says "Goofy wants to be in the leadership to enhance his resume
and told several of us that at a Bar BOF and should not be
selected for _any_ position.  If you don't believe me alone,
check with X, Y, or Z."  In the current model, that would
require writing a separate review for each position.  Lots of
make-work; might not happen at all.   But a Nomcom would really
want that input I think.

(4) I have some basis for believing I can comment on the IAB and
its needs.  Perhaps I'm completely wrong, too old and set in my
ways and that my opinions should be dismissed or negatively
weighted -- that is the Nomcom's decision, not mine.  But one of
my observations is that the IAB's role is not really very well
defined for Nomcom considerations and that the role and utility
of the IAB is strongly shaped by those who are put on it.  In
the interest of full disclosure, if I were [re]inventing the
world, I would significantly reduce the fraction of the IAB that
is appointed by the Nomcom, but that is another discussion.
IMO, the way for me to make useful comments is to explain some
scenarios about what I think the IAB might be like and then tell
the Nomcom which people I would select to advance each of those
scenarios.   With a very small number of exceptions, I can't
evaluate people without those scenarios; simply presenting me
with a list of 20 choices creates high odds of the Nomcom's
getting no comments at all.

Just my opinion, YMMD.

   john

_______________________________________________
Ietf mailing list
Ietf@xxxxxxxx
https://www.ietf.org/mailman/listinfo/ietf


[Index of Archives]     [IETF Annoucements]     [IETF]     [IP Storage]     [Yosemite News]     [Linux SCTP]     [Linux Newbies]     [Fedora Users]