----- Original Message ----- 
From: "Erik van der Poel" <[EMAIL PROTECTED]>
> So, instead, I propose that we start thinking of a "one label, one 
> writing system" rule. The Unicode book defines "writing system" as "a 
> set of rules for using one or more scripts to write a particular language".

This is an intersting idea.  Based on the Unicode book though, there are 6 
defined "writing systems"
- Alphabets (e.g. Latin, Greek, Cyrillic, Armenian, etc.)
- Abjads (e.g. Hebrew, Arabic, Syriac)
- Abugidas (e.g. Devanagari, Bengali, Gurmukhi, Gujarati, etc.)
- Logosyllabaries (e.g. Han)
- Simple Syllabaries (e.g. Cherokee, Hiragana, Katakana, Bopomofo, etc.)
- Featural Syllabaries (e.g. Ethiopic, Canadian Aboriginal Syllabics, Hangul)

And under the 6 writing systems are different "scripts".  It is not difficult, 
I suppose, to try to imagine "fitting" the phrase into a "language" context 
(i.e. as what you have described as a "Japanese writing system"), but I do not 
have enough expertise to say whether that is a good idea or whether it would 
invite other confusions and objections.

Rather than trying to "fit" an existing phrase into the context of IDN, perhaps 
it may be better to just invent a convention to describe what is actually 
happening and plausible given precisely the DNS context.

That is, rather than trying to use the words "language" or "writing system" or 
"script" which invites contention, why not just say that each registered IDN 
Label should have an associated "IDN-Tag".  And an "IDN-Tag" is defined by a 
domain registry and encompasses a policy profile that is developed by a local / 
lingusitic authority with rules relevant for IDN use.  This could include 3 
general types of things:
1. the collection/list of valid characters/codepoints;
2. some contextual rules (if applicable); and,
3. considerations for variant preparations/reservations (if applicable).
(maybe more than 3 would be appropriate, I am just suggesting these as a 
starter)

Today, for example, the .JP registry has created its Japanese policy profile 
with a list of valid characters (which includes kanji, the kana's and LDH) with 
a couple of contextual rules (max length of domain is 15 native characters and 
some defined pre-mapping characters).  The .CN and .TW registries joined forces 
through the CDNC to develop the Chinese policy profile with a list of valid 
characters (which includes Simplified and Traditional Chinese characters as 
well as LDH) and considerations for variant mappings.

Note that both these "Domain-Langauge-Policies" in fact include multiple 
"writing systems" within one profile.  Both also included multiple "scripts" 
and even arguably multiple "languages".

As we continue to work on implementation of IDN, I have found that it is hard 
and often creates confusion when using overly generic words such as "language" 
and "script" because they generally come with a lot of baggage.  In order to 
clarify discussions and try to avoid convoluted debates which may be 
interesting buy ultimately confusing for lingusts and technologists alike, it 
seems that inventing a few phrases to give it a clear "IDN" context would be 
useful.

Just some thoughts...

Edmon

Reply via email to