Re: [gnso-idng] rethinking IDN gTLDs

[Eric Brunner-Williams <ebw@xxxxxxxxxxxxxxxxxxxx>]

Mike Rodenbaugh wrote:
> Thanks Chuck, that helps a bit, but I would like to understand the details
> of how an existing gTLD registry might offer an IDN equivalent "in a way to
> minimize any confusion".  I think I saw a mention of 'sharing a root zone
> file' but there was no explanation.  If this is already explained somewhere,
> then maybe I just need to be pointed in the right direction.

Mike,

The problem was first considered, and several solutions explored, by 
the .cn, .tw, .hk and .ma registries, the Chinese Domain Name 
Consortium (CDNC) in 2000. At the time I worked for NeuStar, and was 
directly involved.
An interesting proposal, interesting to me as a co-author, along with 
Scott Hollenbeck of NetSol, now VGRS, of EPP, was a registration in 
one registry would trigger events in the other, cooperating 
registries. As mechanism, an EPP provisioning client<->server event 
could (this was the protocol part of the discussion) result in one or 
more EPP provisioning server<->server events. The change to EPP was 
not adopted, and provisioning server<->server events were accomplished 
by other means.
[More generally, a change to the EPP client-server model would allow 
registrars to escrow their data with an escrow server, while 
transacting with the registry server, and other things now done out of 
band, and generally manually. End EPP trivia aside.]
My first observation is that intentional similarity is not restricted 
to a single, existing gTLD or other registry, but relies upon plural, 
cooperative agreement.
My second observation is that only a part of any two or more zone 
files need to manifest the cooperation between the zone file managers. 
 The complete equivalence, whether attempted via DNAME or exhaustive, 
string by string equivalences, is not necessary for user benefit in 
the shared script to be preserved by registry operations.
We have unimplemented instances of this in the legacy, ASCII 
registries. A number of cooperatives, and museums, registered their 
names with the only registrar, and registry, accepting 
non-governmental, non-territorial applications at the time. As some 
later point in time, registries for cooperatives and museums were 
authorized by ICANN, and operationalized.
The intent by registries to create non-confusion could have been that 
data of a particular type, such as registrations by cooperatives and 
museums, in one zone file, created data in another zone file.
This hasn't happened, but it could, and you asked, "how an existing 
gTLD registry might offer ...", so I've pointed to the general answer. 
The "... an IDN equivalent" answer Chuck's already offered the "asia" 
and "chung guo" example.
I want to point out, as an example, that while the "views" of 
Arabophone Africans of "Africa" (in Arabic Script) and Francophone and 
Anglophone Africans of "Africa" (in Latin Script) have a non-empty 
intersection, there may be distinct views of a resource associated 
with names, for which distinct resolutions to distinct network 
resources is more appropriate than not.
My third observation is that an approach to the formation of intent by 
registries is necessarily incomplete if the evaluation framework 
excludes user benefit as the motivation of the registry(ies) applicant(s).
And that, unfortunately, is the state of the DAGvX.

Eric