[Date Prev][Date Next] [Thread Prev][Thread Next] [Date Index] [Thread Index]

Re: Adding New Script Variants on Debian Installer



Package: Debian Installer

On Sun, Mar 18, 2012 at 1:38 AM, Christian PERRIER <bubulle@debian.org> wrote:
Quoting Eagle Burkut (eagle.burkut@gmail.com):
> Dear Debian Developers/Maintainers,
>
> How can we have new language/script on debian installer? Is there any
> requirement? If there is, what is the requirement?
>
> Ubuntu uses Debian Installer and it has Arabic based Uyghur. We need Latin
> based Uyghur too.


We need to go though the whole New Language Process as we did for "ug"
with Gheyret Kenji

Please notice however, that a new translation adds an extra size to
the installer so we need some good rock solid rationale for adding a
variant of an existing language.

While we have finished translating more than 95% translation string in Ubuntu (that is 350,000 sentences!), due to lots of bugs and inabilities in various libraries/packages/fonts, we got a localized Ubuntu with lots of bugs/errors. In summary, the original Ubuntu has problems, bugs in every aspect of

(1) Right-to-Left support
(2) Bi-Directional support
(3) Font selection
(4) Proper fonts

Thus, even with hundreds of translator spending couple of years of hard work, we still ended up not so good localized Ubuntu.

If we utilize Latin based Uyghur, all of the above problems will disappear immediately. That is why we are thinking about adding Latin based Uyghur to Ubuntu, including the Debian Installer.

Also, variants have not been proven as working properly as of now : we
for instance have a "sr@latin" variant for Serbian that is not
activated yet because supporting it requires changes to localechooser
that nobody did until now.

(the main problem comes from the fact that localechooser assumes that
the chosen language code and the chosen country can be concatenated :
fr for language then FR as country, give "fr_FR".....while this
doesn't work for "sr@latin" combined with a country code such as "RS"
as the locale should be "sr_RS@latin", not sr@latin_RS)

As Uyghur is written in modified Arabic-Persian in China, in Cyrillic in Kazakhstan and central Asia, and in Latin elsewhere globally, we previously had ug_CN locale in Arabic, and we have just developed ug_US locale in Latin and submitted it to upstream glibc library, and more, we are planning to develop ug_KZ locale in Cyrillic in near future. 


Reply to: