Re: guidance on avoiding too many people from one company/group

[Dave Crocker <dhc@dcrocker.net>]

Donald,


Monday, November 11, 2002, 6:20:53 AM, you wrote:
Donald> I think the main problem is the appearance of fairness.

Appearance of fairness is only one of two concerns. The other is the
legitimate need for real diversity of perspectives. Too much "cultural"
concentration works against this need.

No matter how well intentioned individuals are, we reflect the cultures we
come from.  Companies are cultures.  So are some other associations.

Real diversity ensures that a topic is considered from significantly
different perspectives.  Nomcom deliberations absolutely rely on that
diversity.


Donald> You are waving your hands vaguely when that isn't necessary.

FWIW I meant my comments about diversity to be neither hand-waving nor
vague.  They pertain to a very basic construct in this sort of
decision-making process.


Donald>  The IETF
Donald> is, and is intended to be, a technocratic orgnaization.

Yes, but management is management, and that is a different from technology.
Nomcom is about selecting managers. It therefore needs to pay a lot of
attention to matters that are quite different from technical consideration.


Donald> This is
Donald> the only problem I see, not some vague necessity to be abstractly
Donald> diverse for all dimensions (an obvious impossibility).

Don, I think your reaction is probably representative of our community, but
that is exactly the problem. Most of us are technicians (ie, non-managers).
We are asking mostly non-managers to select mostly non-managers for doing
very difficult management jobs.

As a group we tend to be ignorant of the real skills that are needed to be a
good AD. This is not a matter of intelligence or intent. It is a matter of
skill.

Requiring that the nomcom be really and significantly diverse is not a
matter of abstraction any more than worrying that the nomcom actually know
something about IETF processes. Nomcoms must conduct difficult deliberations
and pursue subtle trade-offs. It absolutely must have a diversity of
perspectives to ensure that essential-but-not-obvious considerations are
raised and that trade-offs are considered robustly.


>> Unfortunately the selection process must rely on a rather small
>> "population", namely folks who self-select to offer to participate. (The
>> pool of candidate nomcom participants is inherently biased, because it is
>> self-selected and because even as many as 100 volunteers makes for
>> statistically problematic pool.) So it is not suprising that -- as we have
>> seen -- the resulting group of nomcom participants is not random enough to
>> ensure community comfort.

Donald> When the nomcom member volunteer pool was around 40, as it was for the
Donald> first several years, this problem didn't occur.

We had a far more homogeneous "population" then.

These sorts of sampling issues are bread and butter in social research
methodology. And, yes, that is the skillset that applies to this discussion
of nomcom selection. How to get a small group that represents a large group
is about sampling methodology.


Donald> frankly, if three
Donald> people from a company ended up on the nomcom but they were all well
Donald> known, decades long participants in the IETF, it would not be much of a
Donald> problem.

The challenge is in writing a reasonably objective rule that captures the
problems but lets the acceptable cases slide.  And, by the way, how would
you feel if EIGHT such people were selected, out of the 10 voting members?


Donald>  It's when some company routinely sends 45 people to IETF
Donald> meetings most of whom are relativley unknown to the community

I mostly agree.  That is why I am inclined to believe we should have two
pools of volunteers.  One with the relatively lax criteria in place today,
and one that is more stringent and requires a history of deeper involvement
in the IETF.


Donald>  and 15 of
Donald> them volunteer and qualify for nomcom and 3+ of those are selected that
Donald> its a problem.

>>     Company affiliation is the one that has surfaced. Are there others?

Donald> It's not "company", it's organization.

ok.


>> 2.  What can we do to make randomization better, in terms of the pool that
>>     members are drawn from and in terms of the mechanism for drawing from
>>     that pool?

Donald> The randomization, if you use something like RFC 2777, is about as
Donald> random as you can get.

Alas, I was afraid that my use of the word randomization would cause a focus
on the algorithm that is used now. However I meant the issue as a macro
issue.

It is really about the methodology of populating nomcom, rather than the
particular algorithm for drawing from a particular pool. The current
algorithm is fine. The problem is size and nature of the pool. The algorithm
can do little about this.


>> 3.  What should be the hard limits, in case the randomization does not work?

Donald> Randomization works by being random.

Not if the pool is too small or otherwise distorted. Again, as I note, this
is not rocket science. Nor is it math science. It is social science, and
there is quite a lot of experience with this problem.


Donald> "IETF management experience" is a whole different matter we should not
Donald> confuse with the issue at hand.

Huh? You mean you do not care how much knowledge about IETF management there
is, among the voting members of the group charged with choosing IETF
managers???

d/
-- 
 Dave Crocker  <mailto:dhc@dcrocker.net>
 TribalWise <http://www.tribalwise.com>
 t +1.408.246.8253; f +1.408.850.1850