Charters, "normal process" versus ISOC, etc. (was: Re

[John C Klensin]

--On Monday, 19 May, 2003 15:05 -0700 "hardie at qualcomm.com" 
<hardie at qualcomm.com> wrote:

> John,
> 	Some responses inline.
>
> At 2:58 PM -0400 5/19/03, John C Klensin wrote:
>> Ted,
>>
>> Let me play devil's advocate for a moment here...
>>
>> --On Monday, 19 May, 2003 10:52 -0700 "hardie at qualcomm.com"
>> <hardie at qualcomm.com> wrote:
>>
>>>> But remember that one of the other problems was the
>>>> shallowness of the pool for IESG positions.  The chosen AD
>>>> would have to sit on all IESG teleconferences and take part
>>>> in those ballots, read the drafts put before the IESG
>>>> etc.etc. too. (Except if these are specifically designated
>>>> out of scope when chartering the temporary AD "position".)
>
> The above was written by Pekka, not me; I was quoting him in my
> note.  Sorry if that was not clear.

I understood that, but the point seemed to be most easily 
responded to in the context of your note.  Sorry for not being 
more clear about it.

>> For whatever it is worth, if we can't find a pool of people,
>> even if  it is small, to fill an IESG position of this type,
>> most of the  problem-statement effort is on the wrong track
>> because real  solutions are impossible unless we can
>> dramatically reduce the IESG  workload or de-skill the AD
>> roles.  I haven't seen proposals, or  even real "problem
>> statements" lately that address either of those  topics.  We
>> are still assuming that, while the system needs tuning,  it
>> is not fundamentally broken.   If we can't find people to
>> serve  on the IESG for each a one-year term, we are
>> "fundamentally broken"  and had better deal with that.
>>
>> I don't think we are in that state, but maybe I'm wrong.
>
> I have seen folks indicating that the IESG role is
> sufficiently demanding that it requires essentially full time
> work, and that the requirement reduces the pool of folks who
> can serve in deleterious ways.

I've heard that.  I believe it is true.   But I note that there 
is nothing in draft-ietf-problem-issue-statement-01.txt that 
even vaguely implies "we need to make decisions in a way that 
sacrifices The Right Thing to Do to the total number of 
qualified people we can get to serve on the IESG at a time".  If 
we think that is a problem, then we had better get it into the 
document.  If we don't think it is a problem, then we shouldn't 
be injecting it into this, or any other, discussion.

As an aside, I note that the closest relevant text is section 
2.5, which talks about the management structure not scaling 
well.  It then goes on to describe the 1992 changes as a scaling 
improvement.   There were any number of good reasons for making 
changes in 1992, but scaling wasn't one of them: by taking a 
model in which one group (the IESG) managed the WG process and 
another one (the IAB) reviewed and approved standards-track 
documents and replacing it with the current model in which the 
IESG has all of those responsibilities, we significantly 
increased IESG workload.  So, if viewed from a scaling 
standpoint along, the 1992 changes were pathological.

>   In this case, I believe we
> could find someone to take on the role, but I believe that
> constraining the choices to the group of folks who can give
> that amount of time  is a big restriction.  As noted below, I
> think the attention the matter is receiving from the community
> as a whole makes who nominally oversees the group less a
> matter of concern than it might otherwise be.  I would
> personally focus the issue on how to get and keep good chairs
> in these roles rather than focus on the appeals process or
> reporting chain.

We may just need to disagree on this.  I suggest...

* One never needs to focus on the appeals process and reporting 
chain if things are guaranteed to go smoothly, if work of one WG 
never overlaps that of another, nothing ever goes wrong, chairs 
don't change jobs or otherwise become inactive, there are no 
disputes about consensus, etc.  In those cases, one barely needs 
an AD except to swiftly process WG output through the IESG.  On 
the other hand, if anything does go wrong, it is not the time to 
start devising procedures since whatever is done is likely to 
stir up more suspicions and unpleasantness.

* We have seen, on this list and in plenaries, a good deal of 
concern and suspicion expressed about the IESG and its ultimate 
commitment to change and openness.  I'm not nearly as concerned 
about that issue as those who have repeated raised it and 
proposed radical alternate procedures as a consequence.  But it 
appears to me that having a process that is as much above 
suspicion as is practical (remembering that I consider anything 
that involves new procedures as impractical) is, at least, a 
good idea.

[... text where we agree deleted ...]

>> And, of course, the IESG and the community could agree --and
>> make  sure the nomcom knows it-- that this particular AD was
>> expected to  take "no objection" positions on any technical
>> documents originating  in other areas.    I don't know
>> whether that is desirable of not,  but it would have a big
>> impact on the workload for that AD and  doesn't require any
>> new procedures (see below).  One couldn't  prevent such an AD
>> from reading all the documents and taking  positions, of
>> course, but, if the question is the total workload,  that
>> option would change the equation.
>
> Less-than-full membership seems to have a whole raft of other
> potential problems which we cannot predict, so I'm not sure
> why this doesn't fall under the same objections you cite below.

Hmm.  I didn't say anything that I would construe as "less than 
full membership".   Unless things have changed recently, the 
typical IESG member, on a typical ballot outside his or her own 
area, records a "no objection" ballot.  Sometimes, that means "I 
read the document really carefully and didn't see any problem", 
but that isn't often the case -- I have observed that, usually, 
if the document has been read that carefully, the person votes 
"yes".  More often, it means, at best, "I glanced through it and 
didn't see any serious problems, so I'm happy leaving the 
decision to those who studied it more carefully".    Now, if 
someone were to always vote "no objection" on everything outside 
his or her area, I'd assume that other IESG members would get on 
his or case about not pulling full weight.  But neither doing 
that, nor not doing that, requires a procedural change -- there 
is no requirement that every IESG member personally review and 
take a position on every document.    And I also haven't 
suggested anything that would reduce the "rights" of that AD in 
any way.  In particular, if he or she analyses a document (on 
_any_ topic) and concludes that there are problems, the 
resulting "discuss" would presumably carry as much weight as any 
other "discuss".

>>> To make a concrete proposal, then, let me suggest that we do
>>> something akin to our normal process but upside-down.  Have
>>> the working group in the General Area, with Harald as AD
>>> tasked with finding the chairs, but have the working group
>>> chairs be confirmed by the NomCom (or a special purpose
>>> NomCom).  This will allow the community to have a voice in
>>> the selection in a way that derives from our current
>>> process. Have any replacement of the chairs require
>>> confirmation by this group as well, so that this oversight
>>> by the community continues.  Given the confidentiality of
>>> that body, normal personnel confidentiality can be
>>> maintained, and those being considered can keep their
>>> privacy until a selection is announced.
>>
>> Ted, I think we need to be very pragmatic about this.
>> Pragmatism  argues for no new procedures if there is any
>> other option.  "The  'General AD' does it" does not require
>> any new procedures.  "Make a  new area with a temporary
>> appointment from within the IESG and a  longer-term one by
>> the Nomcom" requires no new procedures.  In both  cases, the
>> WG(s) are plain, ordinary, WGs, with normal reporting  chains
>> to an AD, which obviously requires no new procedures.  In
>> either case, the AD using an open process, such as the one
>> used to  select this WG's co-chairs, to pick the WG
>> leadership does not  require any new procedures.
>
> Pragmatism would then seem to argue against having a new AD
> for this area having less than complete responsibilities and
> rights, which in turn means we are back to the analysis above
> of the advantages and disadvantages.

You are killing a strawman here.  I didn't suggest anything of 
the sort, merely a little IESG tolerance for a reduced 
out-of-area workload for an AD who was time-limited.  And, of 
course, that sort of tolerance has been applied many times 
before --sometimes happily and sometimes not-- when an AD has 
simply not had the time to do all desired work due to external 
pressures.

>  Pragmatism would also
> seem to suggest against creating a new AD slot for it, on the
> theory that the IETF has never before had an AD with a
> specifically non-technical remit, and that installing one in
> the current system might be a bad idea.

Uh... we actually had an AD whose area/job was "process" for a 
while.  It didn't turn out to be a really good idea (the person 
who held the position is on this list and might want to 
comment), but the mission was different.

> An untidy memory
> reminds me of the arguments as to why there is no Lord High
> Steward as a permanent office in modern Britain, the right to
> hold coronations and try peers being both a temptation and a
> liability for the individual holding it.  Not an issue for us,
> of course, having rejected kings and princes, but you see the
> point.

I would if there were any analogy.  You will recall that I 
suggested this only in the context of a temporary area, with a 
limited duration and an explicit sunset/review date.  I think we 
all recognize the risks of a "process AD" who was inclined to 
create all sorts of process WGs and BOFs to keep the area active 
and busy.  I would hope that the Nomcom would avoid appointing 
anyone with those inclinations.  And, if they developed after 
someone was seated, we would have not only the usual recall 
tools, but a one-year term and a one-year expiration for the 
area as protection.  All within normal procedures.

>> But your suggestion above, which might have considerable
>> merit if we  had infinite time, seems to me to be fraught
>> with requirements for  new procedures and loose ends that
>> should be tied before, rather  than during or after, problems
>> or ambiguous situations.  For  example, the Nomcom has never
>> chosen a WG chair.  Would the  procedures for ADs suffice, or
>> would some tuning be needed?  The  Nomcom also has no
>> procedures for "confirming" a choice made by some  other
>> process.  How much time would the Nomcom spend working
>> through  those issues, and does the framework of 2727 or its
>> I-D successor  provide it sufficient scope/ authority/
>> guidance to do so?
>
> I would argue that we can say: "We will choose N folks from
> among a volunteer pool by verifiably random means to confirm
> that those chairing Working group A and Working Group B are
> individuals appropriate to the task according to the views of
> the community.  If they do not so confirm any individual
> proposed to the task, the relevant Area Director may propose
> one or more replacements if the individual was to serve as a
> single chair and zero or more replacements if the individual
> was to serve jointly."   The issue may be fraught with
> potential issues or loose ends, but getting through it should
> neither take an infinite time nor be impossible.  It may, of
> course, not be worth it. If having the method used by Harald
> to solicit community input on a slate of chairs serves the
> same purpose and seems more open, then let's close the issue
> and be done.  I put it forward as a mechanism that might serve
> should the community's trust of the AD be an issue; if it is
> not, so much the better.

A completely new procedure that, as others have pointed out, 
could easily leave the AD with WG Chairs with whom he can't work 
effectively.  And, if the community doesn't trust the AD, 
chair-appointment is not the only thing the AD does.

>>   Assuming that the WG Chair is selected (or confirmed) by
>>   the  Nomcom, does he or she still serve at the pleasure of
>> the AD, or  does the AD have to initiate a formal recall
>> action to make a  change?  If such a WG Chair resigns for any
>> reason, does the nomcom  have to be reconvened to deal with
>> an interim vacancy, and what  happens to the WG in the
>> interim?
>>...

regards,
    john