Sloppy Charters (was: Re: Discipline of Internet Protocol Engineering)

[John C Klensin]

--On Wednesday, 04 June, 2003 15:08 +0200 Brian E Carpenter 
<brian at hursley.ibm.com> wrote:

> Indeed, scope and non-goals need to be in there too.
>
> The most useful section of diffserv's charter, from the
> viewpoint of discussion management, was the part starting:
>   "The group will not work on: "
>
> I've always regarded the vetting of WG charters as the most
> important single thing that the IAB and IESG do. A good
> charter is the most effective tool an AD or WG Chair can have
> to maintain forward progress.
>
> So, "sloppy charters" needs to be on the problem list.

Brian,

While I fully agree, I think that, in this area and several 
others, we've set up the wrong incentive structure.  When we 
have tight charters, they are often, perhaps typically, not 
enforced.  When an AD enforces one --with regard to benchmarks, 
scope, or procedure-- he or she usually becomes _really_ 
unpopular with the membership of the relevant WG... and most of 
the rest of the community, at best, ignores the action.

Often, we end up with sloppy charters because some group 
insists, very loudly, that _something_ must be done, and done 
quickly, because of perceived market or timing pressures.    ADs 
who try to resist those pressures are criticized as unresponsive 
and foot-dragging, and, in general, no one comes to their 
defense.

We've had proposals whose effect would be to attach real costs 
to sloppy charters and drifting WGs.  While both of the recent 
ones --PACT and the "WG ceiling" draft-- have their own serious 
limitations and problems, they were not even discussed and 
explored seriously (e.g., to look at the problems and figure out 
how to work around them), presumably because people in the 
community don't, in the last analysis, really want those 
constraints on _their_ WGs, regardless of how they might feel 
about those of others.

I'd like to see if, in Vienna and thereafter, we can get ADs to 
report on the draft charters they have turned down, or forced 
into major revisions, because they were sloppy and on the WGs 
they have efficiently shut down for drifting off-charter (or 
just plain drifting).  And then I think we should loudly cheer 
those ADs.

This is of a piece with part of  "WGs produce poor-quality 
work".  Ultimately, poor quality work comes out because the 
community tolerates poor quality work.  The AD, or advisor, or 
random WG participant, who pushes back (either when things are 
under way or "late") is routinely reviled for holding things up, 
causing delays, being opposed to the principle on which the work 
is based, etc.  At plenaries, on mailing lists and, when nomcom 
time comes around, in whispering campaigns, we hear complaints 
about someone who has "caused" a delay; we don't hear complaints 
about those who quietly let garbage go through.  That situation 
requires a careful balance, but, whatever that balance is, we 
clearly haven't found it: the community's position _in practice_ 
is pretty clear.

As an aside, if we do not somehow change this manifest community 
attitude, it will eventually kill SIRS too.  I'd predict that it 
would work well for a while --almost any "change", 
"improvement", or "new method" does-- but that, eventually, we 
will find reviewers who push back on poor quality being abused 
for holding things up that represent clear WG consensus... and 
getting no support for trying to enforce quality.  Some of the 
good ones will decide that they didn't sign up for the abuse and 
drop out, others will start adopting a more relaxed attitude and 
letting things go through... just as many ADs have done.

I suggest that much of the problem here is that, in Pogo's 
immortal words, "we have met the enemy and they are us".   And 
we had better figure out how to turn that around.

       john