[wg-d] comments

[Kent Crispin <kent@songbird.com>]

On Fri, Apr 28, 2000 at 09:44:23AM -0700, Bret Fausett wrote:
> >From my participation with WG-C, I noted that the group often became
> side-tracked by the fact that new members were joining throughout its work.
> It was a distraction, an annoyance, and may have kept the group from
> becoming more cohesive than it might otherwise have been. (I thought it
> worked very well, actually, given its tasks and history.)
> 
> Open membership in WG-C was important for that particular group, coming as
> it did early in the life of ICANN. Last year, it was quite possible that
> someone might have been interested in the creation of new TLDs but ignorant
> about ICANN. I doubt that is the case any longer. We might now want to
> consider freezing the membership of a WG at some defined moment in time.
> 
> The primary purpose would be to create cohesiveness, build community, and
> encourage compromise. With a changing membership, compromise reached one day
> can be attacked the next by a new member not privy to the long conversations
> and debates that went into it.
> 
> Something like a firm 60 day enrollment window at a working group's launch
> would also serve to partially protect WG votes from being stacked.
> 
> On the negative side, such a rule might give a false sense of consensus, by
> pushing criticism to the end (in the public comment phase).
> 
> Any thoughts on this?

I believe that the problems you describe are largely inconsequential
epiphenomema, and that there is a deeper, completely different problem
that destroys the effectiveness of WGs, and will continue to destroy the
effectiveness of WGs until it is dealt with.  To wit: many people are
operating under a fundamentally misguided notion about the "power" of
WGs to decide policy.  Even some members of the NC are confused about
this. 

Consider the following hypothetical:

A group of people interested in some issue (say, the creation of the
".gerbils" TLD) present a charter to the NC that describes the problem
area, and ask that a WG be formed.  The NC notes that the issue is "on
topic" for the DNSO, it meets some minimal standards for coherence, the
charter is well formed, and that the potential WG otherwise meets
minimal formal standards.  The WG is established.  It's charter is to
produce a document -- a formal proposal for the new TLD, in complete and
final detail.  There aren't many people interested in a TLD named
.gerbils, so there are only 10 people in the WG, and they work
feverishly for 6 months, consult legal advisors, contact international
associations of gerbil lovers, talks to large IP interests, works out
UDRP issues, and otherwise cover all their bases; and produce a
definitive, careful, complete document that not only makes a powerful
case for the TLD, but identifies an international non-profit
organization of gerbil lovers that all gerbil lovers respect and that is
willing to accept legal responsibility for contracting with ICANN;
identifies a registry that is willing to run the TLD, etc etc. 

But nobody at all but gerbil lovers participates in the WG.  The WG 
certainly can't claim to represent a consensus of the Internet 
community -- they know that, everybody knows that.  The vast majority 
of the Internet community doesn't even know that the .gerbils TLD is 
being discussed, and could care less even if they did know.

The WG posts its proposal for public comment -- 100 gerbil enthusiasts
send enthusiastic emails in support of the idea.  Finally, the proposal
is presented to the NC, and the NC now must decide whether to forward
the proposal to the Board for final approval.  

On what basis does the NC decide? On the basis of the unanimous and
enthusiastic support for the proposal in the WG itself?

Obviously not.  The WG is composed of 10 self-selected people.

On the basis of 100 enthusiastic email messages? No -- as we all know,
it is just too easy to find 100 enthusiastic people on any given topic. 

While people make much of the "consensus" in WG-C, it is in fact no more
meaningful than the consensus of the gerbil lovers.  The number of
active participants was less than 100, and the active membership of WG-C
is composed mostly of those who have a *strong* emotional investment (if
not monetary) in the pros and cons of new gTLDs.  We can make the claim
that the opinion of two thirds of the <100 active members of WG-C
somehow represents the "consensus" of the Internet community, but on
what basis? Because these people have been involved a long time and have
a strong interest? Because there are members who have views on opposite
extremes who participated?

Step back a moment and look at the composition of the active membership
of WG-C -- less than 100 people we could categorize in various ways:

    N1 people from TM interests
    N2 lunatics
    N3 people from prospective registries
    N4 DNS wars political junkies
    N5 (civil) libertarian idealogues
    N6 gTLD-MoU supporters
    N7 others

It only takes a smidgeon of intellectual integrity to note that the
opinion of such a group on the topic of new gTLDs simply cannot be
considered as a reliable measure of the consensus of the Internet
community.  It doesn't matter what the vote tallies were.  The WG was a
self-selected small group of strong partisans so lost in continual
conflict that they could only barely scratch the surface of the actual
WG charter.  [By the measure of the WG charter, WG-C was without
question an unmitigated failure.  But of course, the charter was out of
touch with reality to begin with.]

It is not necessary to assume a massive conspiracy of TM and large 
business interests to explain the NC's vote on WG-C.

Yet, the NC was in a position where it had to respond to the output of 
WG-C, which, in the grand procession of these events, is real progress.  

Getting back to the situation where the gerbil WG has presented a 
proposal: 

There are several reasonable considerations that might be used to guide 
the vote of a NC member:
- the quality of the proposal itself; 
- some gut feeling about the support the proposal has from all sources; 
- policy implications -- how will a vote for a .gerbils fit with
  current policy; how will it effect future policy;
- support within the individual NC members constituency.

Note that these are indeed reasonable considerations.

A very high quality proposal would of course address all these
considerations and more, and could be very persuasive.  Or it might not
be, and the NC votes to reject it.  Or maybe the NC votes to return it
to the WG because the idea wasn't bad, but the legal basis for the
docuements was weak.  There are many possibilities.  The point is, the
WG proposes, the NC decides. 

Some conclusions I draw from this hypo, and the experience of WG-C 
and others:

1) WGs cannot ever be considered as representative -- not ever.  It is
obvious in the case of WG-gerbil, but it is equally obvious in the case
of wg-c, or wg-b.  Even more important, it is *destructive* to try to
force them into the mold of "representative bodies".  The constant
presumption should be that they are *not* representative; that they are
not intended to be be representative; that the consensus of a WG means
nothing except that the WG members agreed to something; and that part of
their purpose is to make the best possible case for the members pet
cause. 

2) The presumption should be that it is easy to form a WG, but that
failure of the WG is a significant probability.  In the IETF the measure
of success of a WG is the adoption of the proposals by the market; by
that measure IETF WGs have no guarantee whatsoever of success, and the
best guarantee of success is to standardize something that people
already use.  In the DNSO the measure of success of a WG is getting its
proposals passed on to the ICANN Board.  There is no necessity that the 
probability be high.

3) WGs can't work without a significant underlying common purpose that
binds the participants.  A WG with a large minority dedicated to the
failure of the WG is going to fail, regardless.  Therefore, part of the
purpose of the WG model of operation is to segment the debate so that
different groups can get together and construct coherent proposals. 
Concretely, if some group wants to get together and produce policy
documents that make the case for proprietary registries, they should be
allowed to do that in relative peace, by chartering a WG with that as
its explicit goal.  A WG that wants to make the case for non-proprietary
registries should be able to have that as its goal, as well. 

That is, at a very high level one of the measures of goodness for a WG
charter is whether the WG is likely to be able to make headway. 

4) There should be *many* WGs, and they should work with little
supervision from the NC.  The NC should only care about the final
result, and the only committment that the NC should make in the
formation of a WG is that it will vote on the proposals of the WG,
should the WG be coherent enough to form them.  Since the committment 
of the NC is light, approval of a WG by the NC should be relatively pro 
forma. 

5) Therefore, elaborate rules of procedure for WGs are largely a waste
of time: lack of progress due to procedural difficulties is prima facie
evidence of either lack of consensus in the WG, an overly broad charter,
an ineffective chair, or some other structural problem.  When there is a
general split in opinion, there is a mechanism provided that worked
fairly well -- the "alternate proposals" procedure.  This should be
freely used, and augmented to allow for WGs to simply split into
separate WGs. 

6) Voting in WGs, other than informal straw polls to guide discussion,
is largely a waste of time, as well, but if a WG wants to conduct 
votes, that's fine.  The only thing that matters is the documents that 
it produces in the end.

7) Large WGs on major policy issues should be avoided if at all possible.  
Instead, they should be broken into smaller WGs that have well-defined 
advocacy purposes, and the fundamental purpose of the WG should be 
considered as coming up with the best possible arguments for that 
position. 

In Sum:

The thrust of the current WG-D report is towards ever more elaborate and
restrictive procedure, and is in my mind based on a completely
unrealistic idea of what WGs are or can be.  Instead, the procedures
should be lightweight suggestions; in a true "bottom up" manner, WGs
should be easy to form; approval should be pro forma; they should have
few formal restrictions; and WGs should be expected to fail with fair
regularity. 

The early NC set up several monolithic WGs: this was in my opinion a
serious mistake (but they were new on the job, after all, and many other
people fell into the same trap, and are *still* in the same trap.) In
retrospect, the NC should have *invited* the formation of many WGs on
the topics of interest, with the stated goal of collecting coherent,
concrete proposals.  That didn't happen: instead the NC *imposed* in a
strictly top-down manner several WGs, and proposals have been highly
controversial, bitterly contested, and more window dressing than
substance.  

Serious mistakes were made in the beginning; there is no reason to
perpetuate them. 

WGs do not *decide* policy; they cannot *decide* policy; they will never
*decide* policy.  At most they can *propose* policy.  There are many
people who have a strong ideological abhorrence for this fact, but the
recent NC actions should demonstrate fairly clearly that it is indeed a
fact: WGs can propose policy, they can make decisions about what to
propose, but it is ultimately the NC that *decides* what gets forwarded
to the Board. 

-- 
Kent Crispin                               "Do good, and you'll be
kent@songbird.com                           lonesome." -- Mark Twain