Re: URL internationalization!

Masataka,

On Thu, 20 Feb 1997, you wrote:

> Alain;
> 
> > The moral of the story is that we ought to use a worlwide character set
> > where even Japanese characters are integrity-preserved everywhere!
> 
> It has nothing to do with URL internationalization.
> 
> We don't use worldwide characters on our international airline tickets.

The airline reservation systems are extremely old and predate
most of localization or internationalization support currently
available on computers.

A better (not perfect) analogy for internationalized URLs are smail
addresses. How many people in Japan prefer to write a Romaji address
instead of using Kanji? How many people prefer to receive mail
addressed in Romaji (or half-width kana) to mail addresses
using Kanji?


> ASCII-only Latin script is perfectly fine as a formal notation of,
> say, Japanese. For example, Japanese sentence "Doumo" in ASCII
> means "Hi" in English.

A lot of distinctions are lost in many cases when reducing
things to ASCII. In many transcription and transliteration systems,
ambiguity is also introduced. In most cases, the transcription or
transliteration systems themselves are ambiguous, because there
are many of them. Many well-defined such systems make extensive
used of diacritics, for which an ASCII-only variant is not well
defined. A good example is Pinyin for Chinese.
Because there is no well-established way to do the transcription
or transliteration, and because in many cases, it cannot be
done automatically, it is not a solution to the problem of
non-ASCII resource names.


> The current detouring attempt of Unicode/ISO 10646 on local script
> support is that they are confusing people like you with no real
> internationzalization expertise not to be able to distinguish
> "language" and "script".
> 
> So, if you don't want to detour, never say "language-tag" when
> what you need is "script-tag", which is MIME charset.

Criticising other's use of terminology, and then getting it wrong
yourself, is not the best way of participating in a discussion.

MIME "charset" parameter values don't denote scripts, they denote
character encodings. Many "charset"s allow several scripts to be
encoded (iso-2022-jp for example allows Latin, Greek, Cyrillic,
and Kanji/Hiragana/Katakana). For many scripts, there are many
different "charset"s.


Regrads,	Martin.