Re: [idn] Comparisons of the proposals

[James Seng <jseng@pobox.org.sg>]

Paul Hoffman / IMC wrote:
> >But the UTF-5 draft is silent on that issue. Ergo, my "stupid" question.
> 
> 'Twasn't stupid. The draft is under-specified.

Agreed with under-specified. Now, why?

The UTF-5 I-D is *NOT* a draft for IDN. (Paul take note. So comparing UTF-5
I-D to CIDNUC I-D is moot :)))

It is a draft for UTF-5 modeled after the RFC for UTF-8. It specify how you
can encode Unicode characters into a alphanumeric range. And that's all the
I-D does.

The reason is historical that we (when we are doing it at APNG) decide to
split Martin Duerst I-D into two section, one on encodings and one on I18N of
domain names. This is because we realise that I18N of domain names can be done
independent of whatever encodings you use *but* the issues for I18N domain
names remains (almost) the same: normalisation, foldings, interoperability
etc. 

So the answer is, sorry, I have not gotten down to write the second section. I
have promised that like 4months ago but my hands are pretty full with IDN WG
and the requirement doc. 

To answer Bill Semich question, the ASCII TLD can remains as ASCII. Therefore,
.COM remains as .COM, .JP remains as .JP, .KR remains as .KR. There are some
limitations and also some 'switching' and implementation details. I can leak
the cat out of the bag now but I would prefer to do it an I-D. 

-James Seng