>From: Keld Jorn Simonsen [mailto:keld@dkuug.dk]
>Sent: Tuesday, September 26, 2000 3:29 AM
>A draft of 15897 is available at
http://www.dkuug.dk/jtc1/sc22/wg20/docs/n712.pdf
>Locales can be found at http://www.dkuug.dk/i18n/WG15-collection/locales
>These are POSIX ISO/IEC 9945-2 locales.
>Some other locales in ISO DTR 14652 format can be found in the GNU glibc
distribution,
>and possibly also form the li18nux site. (I have not yet looked at these).
>The question in the first hand was only the naming scheme, which is
elaborated
>in 15897.
I actually think that we are closer in agreement than it seems. I agree
that if I want Chinese that I should use zh_CN as a locale.
Where we differ is in use of the variant. You have a single variant for the
locale. I am proposing that the language have a variant as well.
For example if I want specifically the Hakka dialect. I am proposing a
zh-hak_CN locale. This is different from the normal variant because it is a
language variant not a locale variant. It should be done this way because
it iterates better. If I also have zh-hak_TW the country differences are
grouped separately from the language differences. I would take a zh-hak-cn
RFC 1766 code and convert it to zh-hak_CN. If the RCF 1766 code was
zh-cn-hak I would reverse the country language order to zh-hak_CN.
When I look at locales like dk_EU what does it mean to me. I think fr_EU
and wonder what kind of French do they speak? Likewise en_EU.
A couple of weeks ago I was think just like you. I thought that the 15897
style locale system was the greatest thing since sliced bread. But someone
opened my eyes and got me thinking. Maybe we need to separate language from
other resources because so much of the resources are dedicated to language.
We need to reuse can conserve the language resource data as much as
possible. To do that we need a better structure that allows us to better
iterate through the language portions of the locale resources.
This is why I brought the subject up here to discuss.
I do not see how I can use the 15897 locale naming structure to implement a
resource iteration scheme. Nor do I see how to map RFC 1766 names to ISO
15897 locales?
Carl
This archive was generated by hypermail 2.1.2 : Tue Jul 10 2001 - 17:21:14 EDT