Re: Running code

[Robert Elz <kre@munnari.OZ.AU>]

    Date:        Fri, 20 Sep 2002 11:15:35 +0200
    From:        Loa Andersson <loa.andersson@utfors.se>
    Message-ID:  <3D8AE737.9000902@utfors.se>

  | If it is a problem it needs to be fixed, not only in the NomCom,
  | it everywhere.

You're ignoring the difference between the way that the nomcom works
and the way that the rest of the IETF works.

  | how does limit to 2 from the same organization get around the
  | senior - junior issue? Shouldn't it be just 1?

It doesn't, and it could be just 1 perhaps, I suggested 2 (just a
suggestion) to avoid being too limiting, and because 2 of 10 is only
a small fraction, with little overall impact, 3 of 10 is starting to
get close to (practical) control.

  | How do you define "organization"

Yes, that one is a problem, yet to be solved.

  | What happens if one of the members of the NomCom changes job during
  | the period, there by creating a group of 3 from FOO Inc?

Another issue that needs discussion, I have no particular opinion.

  | Are 2 people from a single
  | European country a problem? Are 6 from the US?

No, because it is the perception of who can control what that matters.

  | Are there proof that people from the same organization has acted in
  | a way that has been harmful to the IETF, more than people representing
  | different organizations?

No, no proof at all, not even the suggestion that it might have happened.
Not that we should know anyway, as how the nomcom works is supposed to
be confidential.   That is, if anyone did have such proof, or even a
suspicion, they should keep it to themselves.

But again, what is important here is the perception of fairness.  In the
IETF apart from the nomcom, we can watch, and see several people from one
organisation, all supporting the same idea - and then we can reject it,
or we can accept it, after having evaluated it and received endorsements
form others.  And all of this is in the open.

The issue is that we simply do not (and should not) know what is going on
inside the nomcom, and thus we must do all we can to rid the process of
anything that might even just hint at allowing any form of bias, or
external control.   The random selection is/was supposed to do that, but
if it keeps on producing clumps of appointees from the same organisation
(from what is clearly the same organisation, without needing to worry about
the corner cases), then there is at least the perception of the possibility
of control.

There's another issue as well - when one organisation ends up being in a
dominant position on the nomcom, what is to happen when one of the best
candidates for an open position happens to come from the same organisation?

If the members of the nomcom that are from the same organisation recuse
themselves, then the nomcom is left with (perhaps) way too few members to
make a proper decision (imagine if 6 or 7 of the nomcom were all from the
same organisation).

If they don't, and they appoint the person from the same organisation,
won't there at the very least be suspicions of bias?   You appointed that
person because you know them better than the others, or worse, so you
can take over control of the xyz area...

Then, because of that possibility, isn't it possible that the nomcom would
deliberately not appoint the person from the same organisation as many of
their members, just to avoid those kinds of accusations, or thoughts, of
possible bias?   And might that not mean that the IETF would be left without
the best appointee, just to avoid the appearance of bias?

And if that happens, won't there be people who at least, think, that
the actual appointee wasn't appointed on merit, but only because the
better candidate came from the wrong organisation (the one that is over
represented on the nomcom).

The way to avoid all of this is to prevent the nomcom from being dominated
by any organisation/interest group.

  | I'm all for diversification - if you try to run the argument that we
  | want participation NomCom from as many organizations and background
  | as possible because this is a good thing I would be far more
  | receptive.

No.  That's exactly what we don't want to require.   That is, diversification
is fine, but attempting to specify how to achieve it is impossible.  Much
harder than working out what is an organisation to an acceptable degree
of clarity (because the default can always be, if in doubt, it is the same
organisation - the only effect of that is that someone misses out on being
on the nomcom who otherwise would have been chosen - which for most people
would probably be a relief - but should never be a cause for complaint).

kre