[rfc-i] Some questions on the model and the motivations

Dave CROCKER dhc at dcrocker.net
Thu Jan 6 18:34:32 PST 2011




Olaf,

Thanks for the followup.

Let me preface by saying that I fully believe everyone involved in this topic 
wants it to go well, not the least counting you.  So the criticisms I've been 
offering -- and, alas, that continue below -- are about our community 
performance, not about community or personal intent.


On 1/4/2011 12:43 PM, Olaf Kolkman wrote:
> FWIW I asked the question as 'Olaf (No Hats)'. I realize that the hat is
> never really off, but it was to signal a more personal question.

(This is entirely a side issue, IMO, but I'll engage in the brief distraction:

    1) For something like this, I believe you can't take your hat off; you are 
too central to the topic and the public discussion doesn't really have a life of 
its own.

    2) Worse, I believe you didn't really take your hat off! Your question very 
much came as a management exploration as part of your responsibility as IAB 
Chair. This was a very specialized question coming from a specific management 
interest.)


> More importantly, after reading the thread (I only respond here for context)
> I feel like I should explain my motivation for asking the questions.

Yup.


> On question 1a) The first time the community has been exposed to a plan and a
> motivation was during the Beijing IETF. My impression was that the
> presentation layer at that time was not found sufficient for a content full,
> substantive, and fact based exchange of thoughts.
>
> Shortly before the X-mass holidays the motivation was posted and only then
> the community got a serious opportunity for discussion of the drafts. I

You are asserting a critical-path dependency that I believe does not really exist.

You appear to be saying that a specification and a summary of a specification 
are insufficient for debating the specification, absent a document that explains 
the rationale for the specification.   Since this is not our model for technical 
specs, why must it be true for the RFC Editor "spec"?

An explanation helps, of course, but what I tried to assert was that our 
community has managed to find one excuse after another to keep from doing the 
substance of our serious homework, namely grappling with the substance of this 
topic.

We should be reading, thinking, discussing and converging on the actual work 
that an RSE needs to do and the evaluating the extent to which the proposal text 
does or does not reflect that.  Although there are some notable exceptions, the 
overall pattern of list activity has largely been disconnected postings.


> believe that the IAB will need to make a decision based on having seen the
> community discussion, and after the community understanding the trade-offs.
> In my role as chair I am trying to get towards a process that allows the
> community to chime in and be informed.

For whatever reason, it's clear that the "community" is not able to self-direct 
the public discussion of this topic.

This suggests the need for much more active discussion leadership.

I strongly recommend you and the IAB do whatever is necessary to directly lead 
community review of the options. I think this could be done within a week to 10 
days.

Is something like this is going to happen?


>>      The discussion on this list is, at best, an adjunct to
>> that process; it is not a replacement.
>
> But as argued above, it is a necessity, as it should inform the IAB on
> whether the decision they are about to take will be supported by the
> community. As with any reorganization project one needs a critical mass of
> support in order for the reorganization to be successful.

With few exceptions, the IAB has not been particularly visible for this topic. 
Yet I think it a topic that needs more active IAB participation, not less. The 
IAB should not be a passive body, absorbing our confused community chorus.  IAB 
members should be raising and exploring their concerns directly with the 
community, pressing towards coherent resolution.

Typically, folks charged with making a decision and with soliciting community 
discussion and developing consensus -- put significant effort into ensuring the 
consensus is developed.  For something this important, what we have instead had 
has been oddly laissez-faire.


> On question 1b): I have tried to ask _open_ questions that allowed Glenn to
> motivate the choices and the trade-offs that he made. I would hope that as a
> management consultant Glenn did think about what the tasks are of the RSE and
> if they could have been re-distributed in different ways. His assignment was

I do not recall his being asked to consider this.  It's always a delight to have 
someone do serendipitous tasks that are beyond what they were asked to do, but 
it's not wise to rely on it.

But beyond this, I believe that the body of documents he produced provide 
extensive explanation for the reason the duties should/must NOT be distributed. 
  Nor do I believe that any of that basis is very creative.  In terms of 
management for a critical operation, his choices and explanations strike me as 
thoroughly conservative and well-founded.


> open enough to allow such and I think he has implicitly done so. And frankly
> I believe that 'we've done so for 40 years' is not the best motivation for
> sticking to the model of an RSE as an independent leg in the model that we
> introduced in 5620.

The IETF used to pay some attention to protecting an installed base. Established 
practice used to be a point of concern, in order to protect continued operation 
of a critical resource.  I do not understand the lack of sympathy for that 
concern now.

Stated differently:  I do not understand what seems to be a rather cavalier 
desire to invent a new management structure that is unusual and seems guaranteed 
to have significant, long-term downsides.

(For reference, Andrew has been diligent and clear about his counter-view on 
this. This would be an example of a point in a posting worth pursuing towards 
resolution.  That he seems to consider RFC Editor operations similar to working 
group operations strikes me as problematic, but at least he's offering a view 
with a foundation.  Although he, too, seems distracted by titles, he also 
emphasizes particular work.)


> Maybe I should have asked the question slightly different: Have you
> considered to redistributing the responsibilities that you identified are
> needed for Series Continuity in different ways? What were the trade-offs that
> made you choose for staying close to the  model of 4 legs from RFC5620?

I know this will sound harsh, but as you offer it, that seems a random question, 
one of many possible ones.  Rather than being dropped into the middle of an 
already-confused public discussion, it needs to include careful motivation and 
your own discussion of the reason it appears to be reasonable, desirable, etc.

Major changes to a management model usually carry massive, unexpected and 
undesirable side-effects.  People impose major changes only when deemed 
essential and it is pursued carefully.

In any event, questions have costs.  The first cost needs to be carried by the 
person asking the question.


> That question is not asked because I believe other models work better, but to
> make explicit whether or not they have been considered by Glenn.

Why?  Why is the answer important to know?  How does it help meet the deadline?

Glenn was not asked to produce an academic review of various organizational 
choices.

He was asked to produce a set of pragmatic recommendations.

Different task.

d/
-- 

   Dave Crocker
   Brandenburg InternetWorking
   bbiw.net


More information about the rfc-interest mailing list