From: Peter Constable (petercon@microsoft.com)
Date: Tue May 17 2005 - 13:06:07 CDT
> From: unicode-bounce@unicode.org [mailto:unicode-bounce@unicode.org]
On Behalf
> Of JFC (Jefsey) Morfin
> At 22:22 16/05/2005, Philippe Verdy wrote:
> >If something is wrong here, there's something not documented in
ISO15924
> >(the list of characters that are considered part of the script).
>
> Thank you for this. This is the very point I need to "debug" for
several
> months. A human being can more or less understand "Latn". A computer
needs
> a charset.
This seems to me to be about as obvious as saying that ISO 3166 is
faulty because it's not documented with a set of geospatial vectors
defining the physical extent of countries, and that ISO 639 is faulty
because it doesn't document all of the utterances/texts that make up a
language.
Certainly when it comes to creating implementations you need to know
what characters are considered "Latn", but the fact that ISO 15924 does
not document that does not entail that ISO 15924 is inadequate for its
purpose. In the same way, anybody implementing linguistic processing for
some given language needs to know about valid utterances/texts of that
language, but the fact that ISO 639 doesn't document those doesn't mean
it needs debugging. People find detailed information about particular
languages from sources other than ISO 639; similarly, people can find
details about what characters are associated with "Latn" from sources
other than ISO 15924. As has been mentioned, Unicode happens to provide
information that accomplishes that.
There is *no* need for charsets (using that term in its specific IT i18n
meaning as a standardized coded character set than can be used in
interchange) to accomplish this.
> 5. this way when we quote the tag, we quote the locale. When we have
the
> locale we have all the information you call for (provided the way to
to
> draw letters, but also icones, etc. is included in the locale).
Nobody has *ever* said that a locale tag gives you everything you ever
needed to know to implement support for a locale. It simply functions as
a metadata ID for a given data category; the rich content of that
category is defined separately.
Peter Constable
This archive was generated by hypermail 2.1.5 : Tue May 17 2005 - 13:06:54 CDT