Browser IDN display policy: opinions sought

[Eric Brunner-Williams]

>> The issues between the different options have been discussed at
>> lenght, so let me just say that the biggest problem I have with B is
>> that it leave it up to Firefox to decide what is a good/bad TLD
>> registry. I think that belongs elsewhere, namely with ICANN.
> 
> If we had enough faith in ICANN, that might work. But judging from the
> mails on this list from people close to the ICANN process, I see lots
> of doubts and not much faith at all.

Randy Bush occasionally reminds the NANOG community that "procmail is
your friend".

There are mechanisms, and there are policies.

Whatever the economic dependencies, and therefore the limits on its
policies and choices of mechanism, the dependence of entities engaged
in the development of applications consuming rfc1034/35 et seq.
services is, to a first order, distinct from the dependencies of the
Marina del Rey entity on the Reston entity, and therefore its limits
on policies.

This first order distinction is present in mail transport agents, in
state-full applications, virtualized-circuit applications, etc.

Assuming, for the sake of intellectual interest, that "agency capture"
exists, that the Reston entity determines root zone access policy, and
gTLD policy, more than any other entity, then the recourse to a single
source of policy and the necessary deprecation of all mechanisms
capable of supporting other sources of policy has the effect of
extending the scope of control of a dominating actor in one market to
others.

The act of turning up a nameserver constellation in November 2001 for
which resolution of HAN script labels was neither administratively
failed nor third-party synthesized at a subsequent point in time,
several years before this semantic was available from a pre-existing
nameserver constellation, demonstrates that independent of the
intentions of the IAB of 2000, there are limits to unitary policy.

Perhaps the HTTP application authors will voluntarily adopt a unitary
policy that does not arise from first order dependencies such as
private equity, or public agency capture, and perhaps other
application authors will as well, and perhaps all will adopt the same,
single, unique policy, and perhaps the differences between any two
instances of a zone will not grow without bound, through prudent
stewardship by each cooperating publisher. However, this may not occur.

While it may be feasible to attempt to persuade each application
author set to use a particular source of policy, it is an unfortunate
fact that retail and wholesale monitization schemes are not
distinguished from the public interest. As Peter Dengate-Thrush
responded when I reminded him at the Nairobi Public Meeting that there
is no agreement as to what constitutes "the public interest",
(paraphrase of the original) 'the interests of domainers are a public
interest.'

I honestly don't know if adding brands in many scripts to some zones
or extending the typosquatting "natural traffic" economic model to
more zones or enfranchising indefinitely private parties with limited
resources is in any application author's interest, but given the
plurality of existing views into the union of one or more data sets,
and the expressions of intent to create further views, by governments,
institutional actors, and others who's requirements are not
controlling a particular source of policy, a uni-polar policy
assumption appears tenuous.

My two beads worth,
Eric