[ncdnhc-discuss] ccSO

[Kent Crispin]

On Thu, Dec 20, 2001 at 08:44:13PM -0600, Alejandro Pisanty wrote:
> > > I agree.  I think that the idea of a ccSO is very poorly considered, and
> > > raises all kinds of questions that were very difficult to answer during
> > > the DNSO formation process.  For example, would the ccSO have a
> > > constituency structure? Would there be a NCC in the ccSO? How would
> > > other interests participate? Would there be an IDNHC for the ccSO?
> >
> > Kent, I have some questions.
> > Do ASO and PSO have constituency structure?
> 
> This may not be too relevant for the discussion.

Yes.  The ASO and PSO deal with quite different kinds of issues.  There
is very little end-user involvement in address allocation, and address 
allocation technical concerns are much more obscure than DNS (and of course, 
DNS is pretty obscure to begin with).  The concerns of the PSO has even 
less in common with the DNSO.

On the other hand, there is an enormous amount of overlap between the
concerns of ccTLD users and gTLD users -- in fact the differences within
the classes are as great as the differences between the classes.  (The
differences between .com and .edu users may be greater than the
differences between .com and .co.uk users.)

> ...the NCDNHC was formed to enter dialogue with other types of
> interests in domain names: registries, registrars, ccTLD administrators,
> etc. As long as we are under the same roof we may try to give discussions
> a global significance. The NCDNHC - ccTLD relation is more or less on an
> even footing if the cc's are within the same SO as us non-commercials.
> This balance is broken, to our disadvantage, by an SO, and what I read in
> Kent's questions is that he is perceiving this and asking if there would
> be a way to rebalance this.

I wouldn't say that I'm seeking a way to rebalance this; after thinking
about it for some time I've concluded that, for a variety of reasons, a
ccSO is simply a bad idea -- perhaps not just a bad idea, a *very* bad
idea.  That's not a position I take lightly -- I have some very valued
friends in the ccTLDs. 

The representational issues so far discussed only scratch the surface; 
collectively, I find the problems overwhelming.  They are interrelated,
so it is hard to disentangle them into neat bullets.  But here are some
not-so-neat bullets:

  1) What happens to the DNSO? If a ccSO is formed, you can't just get rid
  of the DNSO, because then there would be no forum for gTLD issues. 
  Chun noted that a gTLD SO would be a bad idea; however, if a ccSO is
  formed, then the current DNSO would in fact become the gTLD SO (gSO). 

  While the structural changes to the DNSO in the conversion from DNSO
  to a (gSO,ccSO) pair might be minor, that isn't the real issue: the
  gSO and the ccSO would have a peculiar relationship, because there
  would be so much overlap in policy issues between them.

  The effects of this on board dynamics could obviously be profound; one
  of the effects certainly would be to move decision making higher up
  the tree -- decisions about DNS policy would in effect be moved to
  negotiations between the upper echelons of the gSO and the ccSO, if
  not to the board itself.  As Milton points out, in the long run the
  most likely effect is to move yet more decision making to the staff

  2) The ccTLD constituency is a constituency of ccTLD registries; it is
  not a constituency of ccTLD users.  It is the ccTLD registries that
  are calling for a ccSO, not ccTLD users.  Without exception, in every
  discussion of this issue that I have witnessed (I have witnessed
  many), the concrete concerns that are expressed as motives for the
  creation of a ccSO are the direct concerns of registries, not users, eg:
  the contracts between ICANN and cc's; how the cc registries should
  contribute to the support of ICANN, modifications to contact data, and
  so on.  [Of course there are *indirect* effects on ccTLD registry
  customers, just as there are indirect effects on the customers of gTLD
  registries.]

  The fact that some ccTLD registries are run by academic institutions
  or non-profits, or even that they have some kind of a governance
  structure, is really irrelevant -- the bottom line is tat allowing
  registries to speak for their customers removes the customers from
  direct effect in ICANN processes.  This is just as true for ccTLDs as
  it is for gTLDs. 

  3) The creation of a (gSO,ccSO) pair will without doubt create
  confusion and disruption in ICANN structure and authority.  This is
  almost certainly work to the benefit of suppliers rather than
  consumers.  As long as there is confusion about authority over DNS
  policy, registries of all kinds will exploit it.  [The net effect is 
  that creation of a ccSO will benefit gTLD registries more than it 
  harms them.]

  4) ICANN may be motivated to form a ccSO because it appears that the 
  cc's will be more likely to deal with ICANN.  I think this is a 
  forlorn hope, though -- the difficulties that ICANN faces in dealing 
  with the cc's stems from the constraints on ICANN from the USG and 
  other realities, not anything under the direct control of the ICANN 
  board.  Giving cc's a seat on the board won't make those realities go 
  away, however much ICANN and the cc's may wish.

  5) IMO, the 4 way split of ASO/PSO/DNSO/AtLarge in ICANN should be
  thought of as primarily a *functional* split, not a *representational*
  split (though of course the distinction is not sharp).  While the
  notion of a functional split is subtle, it does provide a meaningful
  way to divide up the board that doesn't lead to an infinite regress
  of representational structures (ie, the endless creation of yet more
  SOs).  [I should note that I consider the ALSO a special case...] The
  infite creation of SOs is a quite likely result of creating two that 
  are devoted to *representing* special interests.

  As I said, this point is subtle, perhaps better discussed over a beer. 
  But I think it is much more practical to think of the board this way,
  instead of as some kind of legislative body.  It is more consistent
  with a volunteer part-time board that only makes high level policy,
  and leaves most stuff to staff and SOs.  [As a general rule, I think
  that a tendency to force things up to the board level will be a
  constant battle in ICANN -- losers will always seek appeal. 
  Therefore, this is a tendency that must *always* be resisted, because
  there is absolutely no end to it. 

  6) I've already mentioned the overlap between the gSO and the ccSO.  A
  structure with a gSO and a ccSO contains a lot of wasteful
  duplication.  Constant elaboration of ICANN's structure is something 
  that ICANN's critics harp on incessantly; of course, many of those 
  critics are the direct causes of that elaboration.

The cc's as a general class have a interesting position.  On the one hand,
they were delegated through a process outside their respective
governments control, and thus, when dealing with their respective
governments, they use that fact to say "we're private, go away".  On the
other hand, when it comes to dealing with ICANN, they can say "we are
under jurisdiction of our local communities, and thus you can't tell us
what to do".  That is, they avoid public regulation by saying that they
are private, and avoid any kind of private control by saying that they
are under the control of their respective "publics". 

> Bear in mind that the NCDNHC is not only "dot-org", but also dot-edu,
> dot-org.cc, dot-edu.cc, etc. The structure of SLDs etc. vary very widely
> in different cc's.

The NCC, if it were functional, would provide an international forum
where non-commercial interests could find common ground in DNS policy. 
With a ccSO that potential would likely vanish. 

> I would not worry so much about .us as you seem to, in this discussion.

dot-us is in fact an example that argues directly contrary to chun's
point -- the .us situation is an argument *for* a strong ICANN, because
there is essentially no representation of non-commercial interests in
.us. 

[...].

> The internal politics of the cc community are far more complex than this
> and cannot be underestimated.

If anything, that is an understatment!

Kent

-- 
Kent Crispin                               "Be good, and you will be
kent at songbird.com                           lonesome." -- Mark Twain