alternative script naming conventions

alternative script naming conventions

[Mike Frysinger]

[-- Attachment #1: Type: text/plain, Size: 1081 bytes --]

currently glibc uses the convention for alternative scripts:
	<lang>_<territory>@<script>
e.g. we have:
	ks_IN
	ks_IN@devanagari
	nan_TW
	nan_TW@latin
	uz_UZ
	uz_UZ@cyrillic

in bug 4176 [1], a user requested we use ISO 15924 [2] naming
conventions instead (which Unicode/CLDR utilize).  the bug
mentioned the IANA standard [3], but it's been replaced by
ISO 15924 now.
e.g. the script names would be:
	ks_IN@devanagari -> ks_IN@Deva
	nan_TW@latin     -> nan_TW@Latn
	uz_UZ@cyrillic   -> uz_UZ@Cyrl

on one hand, it'd be nice to not invent our own naming (i assume
what we use now came from somewhere, but i don't know where).  it
matters a lot more if we want to expand into more scripts (which
i think we do).  on the other hand, legacy!  we can transition to
the new names and set up aliases for the old to the new so as to
not break existing users.

anyone have an opinion either way ?
-mike

[1] https://sourceware.org/bugzilla/show_bug.cgi?id=4176
[2] http://www.unicode.org/iso15924/iso15924-codes.html
[3] https://www.iana.org/assignments/language-tags/language-tags.xhtml

[-- Attachment #2: Digital signature --]
[-- Type: application/pgp-signature, Size: 819 bytes --]

Re: alternative script naming conventions

[Chris Leonard]

+1 for migrating to ISO 15924 standard 4-letter nomenclature for
script variants.

+1 for making effort to potentially impacted language communities for
additional comment before making final decision

+1 for minimizing impact by maintaining aliases for at least two release cycles

cjl

On Fri, Apr 15, 2016 at 1:55 PM, Mike Frysinger <vapier@gentoo.org> wrote:
> currently glibc uses the convention for alternative scripts:
>         <lang>_<territory>@<script>
> e.g. we have:
>         ks_IN
>         ks_IN@devanagari
>         nan_TW
>         nan_TW@latin
>         uz_UZ
>         uz_UZ@cyrillic
>
> in bug 4176 [1], a user requested we use ISO 15924 [2] naming
> conventions instead (which Unicode/CLDR utilize).  the bug
> mentioned the IANA standard [3], but it's been replaced by
> ISO 15924 now.
> e.g. the script names would be:
>         ks_IN@devanagari -> ks_IN@Deva
>         nan_TW@latin     -> nan_TW@Latn
>         uz_UZ@cyrillic   -> uz_UZ@Cyrl
>
> on one hand, it'd be nice to not invent our own naming (i assume
> what we use now came from somewhere, but i don't know where).  it
> matters a lot more if we want to expand into more scripts (which
> i think we do).  on the other hand, legacy!  we can transition to
> the new names and set up aliases for the old to the new so as to
> not break existing users.
>
> anyone have an opinion either way ?
> -mike
>
> [1] https://sourceware.org/bugzilla/show_bug.cgi?id=4176
> [2] http://www.unicode.org/iso15924/iso15924-codes.html
> [3] https://www.iana.org/assignments/language-tags/language-tags.xhtml