DNSO Mailling lists archives


<<< Chronological Index >>>    <<< Thread Index >>>

FW: [nc-idn] DRAFT MOTION proposal, version 2.

Elisabeth, I am forwarding some of my inquiries to you (provided by
Guillermo, the IPC's Task Force representative) with your initial responses
for everyone's' consideration.

If Verisign is going to be singled out in the resolution, we should also
single out the ccTLDs that have already adopted an IDN system that we
believe is harmful.

Also you indicate that "but the prefix used by this encoding is still to be
determined, and will be at the end of IETF works."  How hard is it for the
testbed like Verisign to change the prefix in the event the one they adopt
is different than the one chosen by IETF?

> A few extra queries from our group:

Why does the resolution only apply to Verisign and not to the cc's?  If the
latter are offering active non-ASCII domain names, isn't that just as
> We would like to see a cite to the twelve things that IETF identified and
> which ones they have accomplished, i.e.:
> "Whereas only an important but insufficient element in the encoding scheme
> has been published to date by the IETF and that element only as a draft"
> Which is this element. This
> element should be specifically mentioned.

==> The chosen encoding scheme is AMC-ACE-Z, but the prefix used by this
    encoding is still to be determined, and will be at the end of IETF
    From maths: you need to be able to implement a two-ways function
    between multilingual representation and ASCII, therefore you need
    to have all possibilities available. Either you do so under a dedicated
    new SLD (and the same accross all TLD - impossible), or you reserve 
    a prefix to be added at the start of ASCII writing of multilingual name.
    For example, in RACE testbed this prefix was "bq--", and then for maths
    validity it was forbidden by rule to register under dot com/org/net
    any name starting with "bq--".
    The prefix for AMC-ACE-Z is NOT yet determined.
> "Whereas there cannot be an open competition at an application level
> all the IDN specifications completed and published"  Why not?  Some ccTLDs
> are already providing IDNs.  

==> It is the same scheme as VeriSign. No application working on global
    Each existing testbed application must assume some encoding
    - they are therefore working localy but not globaly.

    Assume you want to use ML names under .nu, and ML names under
    .com - you will need to have two dedicated customised browsers, 
    because the one for .nu will not decode .com and the opposite.

    You can have multiple people sharing the SAME ASCII-encoded name.
    It will end up with ASCII-encoded A2B7SD9.nu and A2B7SD9.com
    refering to two different names, one in Scandinavian, one in Greek.

> "Whereas it is critical to understand how the whois accessible databases
> IDN would function for gTLD and ccTLD alike"  While we agree with this
> statement, are you saying that Verisign or IETF do not know at this time?
> Have we asked them?

==> I am absolutely sure that IETF and VeriSign know about it. 
    The Universal Whois project, which is at the stage of gathering
    specifications, will certainly give more understanding, and
    hopefully a free software for whois, freely distributed to all 
    Registries and Registrars, and encouraging them to use it.
    If not we will read garbage when interrogating with whois tools
    the ML names from Registrars all over the world.

> "Whereas the International Treaty Organizations, WIPO and ITU, are
> a joint Symposium on Multilingual Domain Names in Geneva,December 6 and 7,
> 2001"  Great, but what does it have to do with rolling out IDNs??  What is
> on the agenda that makes sense for us to wait for?

==> WIPO appreciate that IDN issues are complicated and wants to provide
    forum for education on conflicts.

<<< Chronological Index >>>    <<< Thread Index >>>