Re: Fwd: Eeek - .NU Domains using ヨ, ト , ナ - And More

["Martin J. Duerst" <duerst@w3.org>]

At 10:54 00/01/08 -0500, J. William Semich wrote:
> Clarifying question:
> 
> Isn't the number "2" an ASCII character? Would it not be necessary to mix
> ASCII/numeric characters with Thai, Cyrillic, and Hangul characters? So -
> ASCII characters need to be able to mix with all characters, yes?

Mixing Western Arabic digits such as "2" with many kinds of scripts
definitely brings up a good point. Of course, "2" is an ASCII character.
But in order to avoid this point, I used the term 'script' below.
'Script' is rather well defined for letters, but not in all cases
for digits, and even less for symbols (for these, the term 'pseudo-
script is often used). [sorry to bother everybody with terminology,
but is two up-to-now unrelated areas are comming together, a certain
amount of that is unavoidable]

So as we can mix Western Arabic digits with Latin letters in ASCII-
only labels, I guess we will have to allow mixing them with scripts
such as Greek and Cyrillic, too. Whether we should allow mixtures
with scripts that have their own digits/numbers (Thai,...) is a
slightly different question. The situation is a bit different in
each case, depending on how much traditional digits, and how much
western digits, are used. I guess that rather that making judgement
calls without a clear understanding of the local situation, allowing
the mixture is preferable.

But I don't think that this means that alphabetic ASCII characters
(belonging to the Latin script) should be allowed to be mixed with
other scripts without further consideration. And of course the
various symbols in the ASCII range are an altogether different
story; most of them aren't allowed in current domain names, and
therefore shouldn't be allowed mixed with anything, I guess.


> Are there any similar possibilities required among the non-ASCII
> scripts/characters listed below?

That's what I would like to know, too.

Regards,   Martin.


> Bill Semich
> 
> At 10:22 PM 1/8/00 +0900, Martin J. Duerst wrote:
> >At 13:36 00/01/08 +0800, James Seng wrote:
> >
> >> I think I18N Domain Names must allow mix language characters,
> >> even across different character set. 
> >
> >James - What do you mean with 'character set' in this context?
> >As we are working on requirements, and don't want to discuss
> >too much about encoding yet, it's difficult for me to understand.
> >
> >I agree with you that it has to be possible to mix characters
> >from different languages and scripts in a domain name.
> >
> >What I think we should discuss a bit more is whether it is needed,
> >from an usage viewpoint, to allow to mix different scripts
> >*in a single label*. Let's for a moment exclude the border cases,
> >and take some clear examples:
> >
> >- Is it necessary to be able to mix Thai letters and Hiragana
> >  in a single label?
> >
> >- Is it necessary to be able to mix Cyrillic and Arabic characters
> >  in a single label?
> >
> >- Is it necessary to be able to mix Chinese Hanzi and Hebrew
> >  in a single label?
> >
> >- Is it necessary to be able to mix Korean Hangul and Egyptian
> >  hieroglyphs in a single label? (once the later are defined)
> >
> >What does everybody think?
> >
> >
> >Regards,   Martin.
> >
> >
> >#-#-#  Martin J. Du"rst, World Wide Web Consortium
> >#-#-#  mailto:duerst@w3.org   http://www.w3.org
> >
> Bill Semich
> President and Founder
> .NU Domain Ltd
> http://whats.nu
> bill@mail.nic.nu
> 
> 


#-#-#  Martin J. Du"rst, World Wide Web Consortium
#-#-#  mailto:duerst@w3.org   http://www.w3.org