On Wed, Jan 13, 1999 at 08:23:06PM +1200, Joop Teernstra wrote:
> At 22:17 12/01/99 -0800, Kent Crispin wrote:
>
> >As I said before, constituencies are a method of preventing tyranny
> >of the majority -- there are other alternatives, but they have been
> >considered, and the minorities who stand the most to be affected
> >prefer constituencies.
> >
> >Consider a flat membership model. Suppose there are 10000 members
> >total. Someone proposes that all the expenses of the DNSO should be
> >born solely by NSI. Under a flat membership model, 5001 votes carry
> >the day. Arguably, this isn't fair to NSI.
> >
> Kent and all,
>
> Arguably any majority decision may be unrfair. But at least it will be
> known to be a majority decision, which is an argument in itself.
> Why is it that you think individual members cannot make good and fair
> decisions for themselves?
Because history teaches that they don't. Majority voting is
intrinsically unfair to minorities, and no system will work that
doesn't have some kind of control over the "Tyranny of the
Majority".
> With consituencies, arguably 50 registries, 150 registrars and 10 TM
> interests could outvote
> 1000 DN holders.
But, arguably, the 210 here have *far* more at stake than the 1000:
We vote on whether a gladiator will live or die. It's entertainment
to the crowd; life or death to the gladiator; everybody gets one
vote. As a potential gladiator, would you agree to that decision
procedure? As a registry, would you put your business in a position
where the crowd decides how you run your business? Of course not.
Nobody who has ever run a business would *ever* agree to such an
arrangement, and, believe me, the registries are not.
> >I personally think a "Bill of Rights" model that protects NSI in this
> >case is the best solution. However, it isn't the solution preferred
> >by the groups that have concerns -- eg registrars and registries.
> >And it does have the problem of coming up with the appropriate Bill
> >of Rights.
> >
> A BoR protecting NSI. :-)
> I had not though of a Bill of Rights in those terms. This is why I asked
> you to post your Montererrey Paper on a BoR, so that we can see what our
> idea's of a BoR may have in common.
> Could you please do it, Kent?
It is appended to this message. I used the term "Constitutional
Protections", as the "Bill of Rights" terminology is, I think,
somewhat US-centric.
Note that as things currently stand, the BoR style protections will
probably be built into contracts between ICANN and the registries.
All this stuff on the DNSO is purely a sideshow as far as actual
control over registries is concerned -- the lawyers crafting those
documents will do the real work.
> Or course a BoR is a difficult problem for those who do not like to concede
> rights akin to property rights to DN holders.
> Yet, investment in Web commerce demands that DN holders cannot be left to
> the whims of registries.
You are lost in hyperspace without a paddle here. The issue of
property rights for DN holders is completely orthogonal to the issue
of a BoR, and can be argued from either side.
[...]
> Yes, everyone would try to secure what they could. But it is better to
> discuss these rights up front than at a stage when the vested interests
> have already rigged the structure of the DNSO in their favour.
Since I wrote the appended paper, I have give a fair amount of
thought to the question of how a BoR would be developed. I don't
have time to develop the argument now, but it would be a *lengthy*
process, and it is important to keep one's eye on the ball: the goal
here is to introduce fair competition in the dns registration
business. A constituency based system isn't ideal, by a long shot.
But, given the constraints, it isn't bad.
> >Consider also that some constituencies, namely the registries and
> >registrars, really are in a special relationship to ICANN in this --
> >they will be the ones most heavily impacted by by ICANN policies;
> >they undoubtedly will be the ones under some form of legal
> >obligation to ICANN.
> >
> So? Does this mean that they need a voting advantage over the DN holders
> who are equally impacted by ICANN policies but who are even more impacted
> by registries policies?
Initially, ICANN will have very little control over registry
policies. It may grow with time, but only if a relationship with
registries is established to begin with.
>From the registries perspective your position is that the users
should have absolute control over their businesses -- that is the
direct implication of a flat representation model. No business
person would agree to that, unless at the point of a gun. ICANN
doesn't have a gun, and neither do you.
Here's the paper:
================================================================
Constitutional Protections and Representation
Kent Crispin
Nov 15, 1998
At the Barcelona meeting Bernard T. (representing ccTLD interests)
proposed the following numbers for representation of constituencies
on the Governing Committee:
Registries: 8
Registrars: 3
Network connectivity etc: 3
Commercial users: 2
Trademark interests: 0 (ex officio standing committee)
At large: 3
The primary motivation for the obvious unbalance expressed in
Bernard's proposal, as he expressed it, is a fear that the competing
interests would essentially gang up on the registries. That is, (to
put it in egocentric terms) the intent of having such a large
representation is not my desire to control others, but rather my fear
what the others might do to me.
The problem identified here is commonly known as "the tyranny of the
majority". It is the major problem with popular democracy.
Bernard said that an alternative that would perhaps be acceptable
would be to have a smaller representation, but to also have some kind
guarantees built in that that would prevent the majority from
running roughshod over registry interests. These guarantees might
be called "constitutional protections (CPs)".
There are several substantial advantages to a CP model. CPs are:
- Well-modulated: CPs can be constructed to exactly fit the issues
of interest to a particular constituency.
- Not a zero sum game: protecting constituencies by adjusting
representation is intrinsically zero-sum -- every representational
unit added is a unit subtracted from someone else. Adding
constitutional protections for one group, on the other hand, doesn't
necessarly reduce the power of another group.
- Flexible: There are mechanisms for changing CPs -- super
majority votes, and so on. These mechanisms are much more
difficult to engage than usual decision making processes, and are
seldom employed. But they are available in the case where there
is very widespread agreement that something needs to be changed.
There are a couple or relatively minor disadvantages to a CP model.
- First, it enables what has been called "the tyranny of the
minority", where, through the power granted by a CP, a minority
blocks some change that a large majority wants. For example, if,
through a CP, a constituency were given a veto power on budget
decisions, that constituency could hold the rest of the
organization hostage and get its way in matters not related to
budgets. This problem can sometimes be dealt with by careful
construction of the CPs, but probably cannot be totally eliminated.
- Second, a CP model adds complexity to the decision process --
perhaps a lot of complexity. Without a CP you take a vote, and if
the vote meets the quorum and majority requirements, the decision
is made. But CP's require enforcement and interpretation, and this
additional mechanism could be considerable.
On balance, however, the advantages seem to me to far outweigh the
disadvantages, so I would like to propose that we work from the
premise that we will have a CP model. More than that, I would argue
that some form of CP model is actually inevitable, and we would
probably be better off starting with that explicitly in mind. (*)
Presuming that we start from a CP model, I would further propose that
we adopt a "constituency-blind" representational model, where every
constituency gets the same number of seats.
Concretely:
Registries: 2
Registrars: 2
Network connectivity etc: 2
Commercial users: 2
Trademark interests: 2
At large: 9
Each of the 5 defined constituencies will potentially have a set of
CPs associated with them -- certain issues about which they have
special powers or protections. The apparent unbalance with the At
Large group is really an illusion -- the At Large group represents
every other possible constituency, but at the same time has NO
specific CPs (**) associated with it -- it can't, because, by
definition, it doesn't represent any *particular* constituency.
Furthermore, the At Large group is large enough so that meaningful
geographical constraints can be applied.
This model hinges on the Constitutional Protections involved. If we
decide to go this route, the debate should shift from the "number of
representatives" to "what special rights and powers should be assigned
to each constituency".
This would, in my opinion, be a healthy development. It would force
us to look at specific policy areas and concerns, and take us away
from thinking in terms of power politics.
-------------------------
(*) It is inevitable because, when you think about it, very
unbalanced representational models create a strong force for some
kind of guarantees for the rights of those not so well represented.
(**) There may be general CPs that apply to every member, regardless
of further membership in a constituency.
--
Kent Crispin, PAB Chair "Do good, and you'll be
[EMAIL PROTECTED] lonesome." -- Mark Twain
__________________________________________________
To receive the digest version instead, send a
blank email to [EMAIL PROTECTED]
To SUBSCRIBE forward this message to:
[EMAIL PROTECTED]
To UNSUBSCRIBE, forward this message to:
[EMAIL PROTECTED]
Problems/suggestions regarding this list? Email [EMAIL PROTECTED]
___END____________________________________________
