I-D ACTION:draft-ietf-problem-process-00.txt

[John C Klensin]

--On Friday, 16 May, 2003 17:13 -0700 Dave Crocker 
<dhc at dcrocker.net> wrote:

> At base, the reasons do not matter.  It simply is not
> reasonable for any one (or two) people to weild that much
> power.  A group based on Rough Consensus should rely on those
> having a point of view being able to recruit others to support
> that view.  If they cannot succeed in their recruitment
> effort, there is something wrong with their view.
>
> We believe in technical expertise, not technical deities.
>
> Expertise is compelling, when the expertise is real. Only
> deities warrant individual veto authority in this community.

Dave, I agree with all of this, except that there is a conflict 
of roles (if not of interest) built into the process.  I believe 
that we either need to eliminate that conflict (which would, I 
suspect, require some really radical changes) or that we need to 
be very careful about how far we go in the direction you (and 
the document) are suggesting.

Explanation...

We expect an AD to take a the lead role in creating a WG, to 
appoint its leadership, and to maintain oversight management on 
its work.  We also expect that AD to carry the WG's work into 
the IESG and to strongly advocate the WG's results.  If I 
remember correctly, current IESG procedures won't permit a WG 
document to advance unless the AD responsible for that WG is on 
record as favoring it.  The portion of the community who believe 
that a WG should be able to determine what becomes a standard 
and whether issues raised from other perspectives are important 
enough to justify revisions would presumably consider a 
responsible AD to be derelict in his or her duties if there 
wasn't strong advocacy for the WG position.

That sets up a poor environment for prohibiting the type of 
substantive "even if this is fine from the perspective of X, it 
will wreck global routing and those issues need to be at least 
addressed" stopping rules that Brian is advocating.

I think that there are several other, less risky, solutions for 
the problem you are concerned about.   The first, and most 
important, is much more openness about what is happening -- any 
action within the IESG that is likely to produce a long delay 
should be required to be accompanied by a public explanation 
(see my suggestion on the Poised list about IESG Charter 
language).  The IESG should be able to approve more time for 
processing a document --or convincing each other-- when 
specific, but not-yet-convincing, objections are raised from one 
area about a proposal from another, but that approval should be 
accompanied by a public statement about what is going on and 
what the objection-placeholder text is.

Finally, procedurally, I don't think we have a mechanism for a 
veto today so there really shouldn't be a discussion about ways 
to eliminate that mechanism.  What we have instead are 
circumstances in which there is a potential for a de facto veto 
when procedures are not followed and a tolerance for that 
behavior based partially on an IESG dynamic of "we have to work 
together tomorrow regardless of what happens to this document, 
and therefore must defer when one of us feels strongly".   That 
dynamic is natural, and we shouldn't try to change it 
significantly -- but it can also be used to cover up, or even 
encourage, well-meaning but pathological behavior... and we need 
to explore ways to stop that set of cycles.

     john