ietf
[Top] [All Lists]

Re: Myths of the IESG: Reading documents is the problem

2005-08-09 13:33:59

--On Tuesday, 09 August, 2005 12:04 -0700 Dave Crocker
<dhc2(_at_)dcrocker(_dot_)net> wrote:

There seems to be a common perception that a major time sync
for the IESG is reviewing documents and writing up comments.

Sam, thanks for pursuing a line of discussion about where AD
time is spent.

You've used the same word that I heard in the plenary, namely
"reviewing". I'm thinking that it was well-chosen, but might
be more complicated than it first seemed.

While it's good to hear that the basic act of reading and
commenting is not necessarily the major time sink that one
might have thought, perhaps the deeper problem is the
difference between the role of facilitating IETF efforts
versus the role of acting like a technical contributor.  That
is, Process Assistant versus Technical Expert. Both tasks are
difficult and both are usually necessary.  And the former also
requires reasonably good technical skills.  But it is very
different, indeed, from the latter.
 
So perhaps the deeper issue with "reviewing" is that Area
Directors might be spending too much time as technical
contributors and not enough helping working group chairs and
editors ensure timely, solid progress.
...

Dave,

I think we agree, but parts of your argument and the distinction
you are trying to make confused me.  Let me try to make a
slightly different set of distinctions and see if we are
reasonably in agreement.  And let me try to state it extremely,
more extremely than may be plausible in practice.

The difference is less between "process assistant" and
"technical contributor" than it is about timing and issues of
reviewing/ approving one's own work:

Once a document appears for final "approval" review, if
difficulties are found with the specification, that indicates a
systemic failure.  That failure was described during the Problem
Statement work (including, I think, by you) as "late surprise".
But it is worse than that because such late surprises typically,
perhaps inevitably, turn the end game of the document approval
process into a negotiation between a WG that thought it was
finished and the IESG about what will be accepted.  That is,
under the best of circumstances, a lousy way to do
engineering... even when it is unavoidable.

If an AD, or a handy IAB member, walks into a WG meeting while a
document is still under development and says "this is going to
need to address the following issue or it is going nowhere", I
think that is great --and your example of EKR and Russ making
that sort of statement in and around a BOF is an excellent one".
But I would not condemn an AD who says, at a similarly early
stage, "please consider the following alternative, which I am
going to outline in technical detail".  While the substantive
difference is clear in some cases, in others, the line is very
faint... and not one that, IMO, appears to be worth trying to
make more precise.  After all, the ADs do have considerable
technical expertise and we should be anxious to take advantage
of it.   The differences between that type of early,
alternative-presenting, technical contribution and the variety
with which I, at least, have problems are, precisely, that many
of the current type come late and appear as "make me/us happy or
you will be stalled forever" commandments, not as alternatives
to be considered early on.

The second problem is that these "late surprises" cost us time
and, I believe, quality.  Whether the time-cost is in reviewing
documents or in the subsequent negotiation and iterations
(remember that, according to Bill's latest figures, 75% of
documents submitted to the IESG get at least one DISCUSS that
has to be sorted out, often without the quality of consideration
that would go with a WG including the relevant issues in its
normal development process.

And the third one is the case in which an AD is not _a_
contributor to the work of a WG, but becomes the (or a) primary
source of technical input to the WG.  The problem is, I think
obviously, much more severe if the AD input appears late (e.g.,
just before or after Last Call) than if there is more
opportunity for the WG to work out other solutions if
appropriate.  And the notion of an AD who has contributed
technically to a WG in some significant way then pushing back
during IESG review if the WG reaches some other conclusion is
pretty close to intolerable.  Changing the review model would,
presumably, clear that situation up in a hurry.

     john


_______________________________________________
Ietf mailing list
Ietf(_at_)ietf(_dot_)org
https://www1.ietf.org/mailman/listinfo/ietf



<Prev in Thread] Current Thread [Next in Thread>