RE: Urdu IDNs: Characters in domain names

This is a good interim solution, which we are also currently aiming to
explore through PAN Localization project (www.PANL10n.net
<http://www.panl10n.net/> ).  We are creating aliases for ccTLDs for Pashto,
Bangla, Dzongkha, Khmer, Lao, Mongolian, Nepali, Urdu and Sinhala and trying
to look at the challenges associated with the solution.  We would like to
hear your comments on the following as well:

 

1.	What is the possibility of including this local language ccTLD
information within the locale definition of a language?  This would also
address valid concerns raised by Richard, as that would ensure that ccTLDs
are compatible with the characters on the keyboard.
2.	What about the gTLDs?  This may not be a parallel set across
languages and cultures and thus simple aliasing may not work.

 

Regards,
Sarmad

 

  _____  

From: mark.edward.davis@gmail.com [mailto:mark.edward.davis@gmail.com] On
Behalf Of Mark Davis
Sent: Friday, August 24, 2007 2:51 AM
To: md@ictenterprise.co.uk
Cc: Richard Ishida; Martin Duerst; Sarmad Hussain; Jonathan Rosenne;
www-international@w3.org; public-iri@w3.org; psayo@idrc.org.in; Maria Ng Lee
Hoon; nayyara.karamat -; cc@panl10n.net
Subject: Re: Urdu IDNs: Characters in domain names

 

I disagree; that strikes me as impractical.

Fundamentally organizations are going to want TLDs based on their view of
the utility of those TLDs to the majority of users. I can well imagine that
a TLD for Iran might use a character that is not in Arabic; by analogy a TLD
for Austria might reasonably use RÖ (Republik Österreich). 

What would work is if the TLDs are always treated as equivalents (aliases)
in DNS lookup; that way Joe Smith looking at a tourist site could always
type "abc.au" and get the same results as if he had typed " abc.rö".

Mark

On 8/23/07, Debbie Garside <md@ictenterprise.co.uk> wrote:


Richard wrote:

I think that, if we are to use non-latin
> characters for script-based TLDs, they must only be
> characters that are readily accessible from keyboards of
> people writing any language that uses that script. 

I agree - insofar as is possible.

Best regards

Debbie Garside



> -----Original Message-----
> From: www-international-request@w3.org 
> [mailto:www-international-request@w3.org] On Behalf Of Richard Ishida
> Sent: 23 August 2007 19:19
> To: 'Martin Duerst'; 'Sarmad Hussain' 
> Cc: 'Jonathan Rosenne'; www-international@w3.org;
> public-iri@w3.org; psayo@idrc.org.in  <mailto:psayo@idrc.org.in> ; 'Maria
Ng Lee Hoon';
> 'nayyara.karamat -'; cc@panl10n.net
> Subject: RE: Urdu IDNs: Characters in domain names
>
>
> > -----Original Message----- 
> > From: Martin Duerst [mailto:duerst@it.aoyama.ac.jp]
> > Sent: 20 August 2007 07:37
>
> > As an example, consider the TLD for Switzerland, "ch". 
> > Switzerland is a multilingual country with four official languages
> > (see top left of http://www.admin.ch/).
> > It would only be confusing both inside Switzerland as well 
> as outside
> > if different languages used differnt TLDs for Switzerland. For many
> > people, the "ch" is just conventional, best known because
> it appears
> > on the back of many cars. 
> > The "ch" is actually taken from the Latin (language, not
> > script) name of the country, "Confoederatio Helvetica", but many
> > people don't realize that, and for TLDs, it doesn't really matter. 
> > What matters is that people who want to know the TLD of Switzerland
> > can look it up, can remember it, can type it, and so on. It's a
> > benefit if a TLD is easily derivable from the country name 
> (e.g. "fr"
> > for France), but it's not always so, because otherwise,
> there would be
> > clashes. It would be very confusing if a TLD changed depending on
> > language ( e.g. "ge" for Germany in English rather than the current
> > "de" (Deutschland, Germany in German), or "al" for
> Allemagne (Germany
> > in French), or the many other names that Germany has in various 
> > languages. It might help some people a tiny bit, but it
> would make it
> > impossible to send URIs using these TLDs across language
> boundaries,
> > and would lead to conflicts because there are only so many 
> two-letter
> > combinations.
>
> I think it's important to note that this only works well
> because people writing any of the Swiss languages or English
> can easily type the letters 'ch' from their keyboard.  If the 
> TLD had been ch I think there would have been a lot of
> problems.  I think that, if we are to use non-latin
> characters for script-based TLDs, they must only be
> characters that are readily accessible from keyboards of 
> people writing any language that uses that script.
>
>
> RI
>
>
>
>
>









-- 
Mark 

Received on Friday, 24 August 2007 14:17:34 UTC