Re: General internationalization

[Paul Hoffman / IMC <phoffman@imc.org>]

At 05:33 PM 1/18/00 +0900, Martin J. Duerst wrote:
>At 16:20 00/01/17 +0800, James Seng wrote:
> > Paul Hoffman / IMC wrote:
> > >      Implementation must specify what character
> > >      sets are used and how these characters are encoded in the DNS names
> > >      and records.
> > > This is bad. I strongly disagree that there is a *requirement* for
> > > multiple character sets or encodings. Those who feel that this is a
> > > requirement should say why it is so. I believe that adding "All" to the
> > > beginning of the first sentence is sufficient.
> >
> > Ok, i think it is bad phrasing. I do not mean there is a requirement for
> > multiple character sets or encodings. What I intent to mean is there should
> > not be a requirement to reject multiple character sets. However, each
> > character sets are defined properly if multiple are used.
>
>Whether to use a single or multiple encodings is a decision that
>will have to be taken at a later stage. I read the above as
>trying to say: Whatever character sets/encodings (one or many)
>we are going to use, it must always be cristal clear which one
>is in use, for each data item.

Assuming that is what James meant, I am in full agreement. But it needs to 
be said more clearly than in his first draft. How about:

The encoding or encodings used by the client in the internationalization 
must be specified in the protocol, and they must be unambiguously 
understood by the resolver.

>Questions of language tagging correctly put aside, I think there have
>been clear objections to require that there is a need to mix scripts
>in labels. In the discussion, we have found a few cases where we
>know we want to allow them (numerals and everything else; mixed
>Kana/Kanji for Japanese), but the general case is not at all clear;
>just saying 'why not' doesn't make this a requirement.

I'm confused here. You said that there are objections, but then you list 
cases that refute those objections. If there are cases of potential domain 
names that need mixed scripts, then we must say that they at least SHOULD 
be allowed by any solution that meets the requirements.

Also, please let's remember one of the main uses for domain names: for 
companies and products. I have seen many examples of company names that mix 
kanji or katakana with Latin characters on billboards in Japan, and am 
pretty sure that I have seen many examples of similar company names  from 
India.

--Paul Hoffman, Director
--Internet Mail Consortium