From: Ruszlan Gaszanov (ruszlan@ather.net)
Date: Sat Jan 03 2009 - 06:08:50 CST
Asmus Freytag wrote
>> How about using Unicode tags on PUA runs (like language tags) to specify >> PUA convention used?
>>
>How about learning from the deprecation of the language tags and from
>ISO2022 that using stateful controls in a plain text environment is a
>bad idea?
>
>Attributes for runs of any kind are best handled by protocols that
>already are organized to recognize text runs. They don't belong at the
>character encoding level.
Yes, we know that in the ideal world stateful controls a bad. But it in the real world (as the introduction of language tags has proven) we still need them sometimes because a stateful encoding mechanism for some types of information is better then no encoding mechanism at all.
I think PUA-convention ID tags are even more necessary then language ID tags, since the information about the specific PUA convention used is much more critical for correct interpretation of the text then language information. Great many users who need to rely on specific PUA conventions for exchanging their text data would benefit from such mechanism because this will enable automated systems (like search engines) to handle PUA codepoints. Even though this solution is far from ideal, this could be a flexible open-end ad-hoc solution for scripts not yet encoded - the mechanism is already there with language tags, all that needs to be done is assigning one codepoint that will identify the tag as PUA convention ID.
Ruszlán
This archive was generated by hypermail 2.1.5 : Sat Jan 03 2009 - 06:11:10 CST