[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: [idn] Determining equivalence in Unicode DNS names

Hi Soobok Lee:
       Right !  Registry-defined equivalence will not do much help to TAG Domain Name in content of web page and mail .  The Localized Equivalence of shared CJK UNICODE may  produce more troubles in gTLD IDN. 
 
L.M.Tseng
----- Original Message -----
From: Soobok Lee
Sent: Tuesday, January 15, 2002 4:44 PM
Subject: Re: [idn] Determining equivalence in Unicode DNS names

----- Original Message ----- 
From: "Stuart Cheshire" 
> 
> Hence my question was whether DNAME might provide an answer, to free IDN 
> from the equivalence debate, by allowing equivalence to be determined 
> locally on a per-zone basis.
> 
 
Applications often try to match labels without making any queries on 
authoritative DNS servers. eg) certificate verifications.

ie. Registry-defined equivalence in such a way cannot be reflected in 
applications' own comparisons of iDN labels. That would cause confusions.
** moreover, some folks are considering localization ([vendor+TLD]-specific preprocessing)
and multiple registrations as a compensation  for incomplete stringprep equivalence rules. But such tricks
would also cause similar confusions and therefore would hurt uniform interpretation of  IDN identifiers 
across different  user environments/applications,  as  was warned by RFC2825 (A tangled web).
 
Soobok Lee
However, directory/search approaches including your DNAME proposal are worth 
to consider. My personal preferrence is :  "<ML>.tld" as keywords/handles
in directory/search engines operated by TLD registries which can
employ their own equivalence rules and provide with advanced features
like fuzzy matches and ambiguity resolutions and more interactivity.

In coming IRNSS WG, I wish we could talk more about directory approaches
that will compete with or replace or obsolete IDN.

Soobok Lee