Re: [ietf-provreg] EPP and IDN

["Edmon Chung" <edmon@neteka.com>]

Hi Klaus,

As Scott mentioned, I think you are right too.

My draft (draft-chung-idnop-epp-idn-00.txt) was created for registries who
are interested in handling character equivalency issues such as trad/simp
Chinese or Latin based character equivalencies.  The extension intends to
allow registries to announce character equivalency policies and
manageability as well as allow registrars/clients to further provision for
parameters within a set/bundle/package.

The concepts laid out in the draft I believe is currently tested at TWNIC's
IDN/EPP testbed.  Also, I will be updating the draft to -01 in a short while
with some corrections and also options on handling cases where large volumes
of equivalent domain sets are created by charprep policies.  The .BZ IDN
registry and perhaps the .SG registry (whom are still deciding on their
policies for IDN) will also try out the extension.

I understand that this is not currently a WG draft, but it would be great if
I could get some feedback and comments on it. :-)

Edmon



----- Original Message -----
From: "Hollenbeck, Scott" <shollenbeck@verisign.com>
To: "'Klaus Malorny'" <Klaus.Malorny@knipp.de>; <ietf-provreg@cafax.se>
Sent: Thursday, July 31, 2003 7:14 AM
Subject: RE: [ietf-provreg] EPP and IDN


> > can anyone tell me something about the relation between EPP
> > and IDN? I
> > have seen that there are various drafts out there (like
> > draft-chung-idnop-epp-idn-00.txt) that deal with IDN issues like
> > variants et al. But is my assumption correct that the base protocol
> > itself does not limit the use of internationalized names in
> > domains and
> > host objects? I mean since XML is able to transport the full Unicode
> > character set, it should be possible to specify non-ASCII
> > characters in
> > domainname related elements. Or does the protocol require that those
> > names are transmitted in their Punycode equivalent, as it happens
> > (AFAIK) in Verisign's current RRP implementation? Is the
> > choice between
> > the "presentation form" and Punycode at descretion of the
> > implementing
> > registry?
>
> I think you've got it right, Klaus.  I'm finding that some communities and
> registries want more information than what is carried in the protocol
> currently, such as language tags and name bundles.  Those features can be
> added with extensions as described in drafts like Edmon's.
>
> The base protocol can handle any character encoding supported by XML.
Some
> registries like to work with the punycode representation to check for
> conversion errors as the name moves along the provisioning path, but
that's
> not a protocol requirement.
>
> -Scott-
>