chinese case in current languagechooser (was Re: Languagechooser changes)

["Carlos Z.F. Liu" <carlos_liu@yahoo.com>]

On Mon, 9 Feb 2004 07:40:30 +0100
Christian Perrier <bubulle@kheops.frmug.org> wrote:

> Special cases:
> 
> -Chinese: up to now, zh_CN was designed as "Chinese (Simplified)". I
>  voluntarily changed this to "Chinese (China)" as this is what the
>  zh_CN locale really means. In the same time, I changed "Chinese
>  Traditional" to "Chinese (Taiwan)" as zh_TW means this.
>  Carlos, translations should be adapted accordingly
> 
>  I insist on this: as long as Chinese will only have one ISO 639
>  code, the variants need to be named after country names (same for
>  Portuguese, btw)
> 
>  If two languages are completely different, they shouldn't share the
>  same ISO code (Chinese and Portuguese are imho ISO anomalies on that
>  matter)

Thanks for your work on languagechooser's improvement, but the current situation
of chinese case isn't acceptable. To let you understand the reason, I will
present some basic knowledge and background about Chinese.

 BACKGROUND:

  Unlike most western language, Chinese is a complicated hieroglyphic language.
  This fact results about 3000 ~ 5000 frequently used characters, and some
  different speaking methods.

  People in north china and taiwan speak chinese as mandarin; people in south
  china and HK speak it as Cantonese.

  In writing aspect, Chinese was developed in two branchs in last 50 years,
  because of some historical, political and regional reasons. They are:

   1) Traditional Chinese
   As the meaning of its name, it's the only Chinese language in the world
   before 1950s. It's still used in Taiwan, HongKong and Macau now.
   locale: zh_TW, zh_HK
   charset: Big5, BIG5-HKSCS (BIG5-HKSCS is a BIG5's extention which add some
            special characters for Cantonese)

   2) Simplified Chinese
   It based on Traditional Chinese and was developed in chinese mainland since
   1950s to make learning process easier. It simplified some characters' shapes
   and merged some characters together to reduce the number of characters that
   were often used. Nowadays, China and Singapore use it as the official
   language.
   locale: zh_CN (zh_SG wasn't supported yet, but you still can find a file
           zh_SG in /usr/share/i18n/locales/)
   charset: GB2312, GBK, GB18030

You can easily find the differences between Traditional Chinese and Simplified
Chinese in languagechooser's interface. Apparently, some characters are the
same, and some are not.

In current languagechooser, "zh_CN(China)" just maps to the Simplified Chinese
interface, and "zh_TW(Taiwan)" means Traditional Chinese interface. But when
users choose "zh(Other)"，the installer couldn't decide which kind of chinese
should be shown. That's not what we want. If we must set a value to "zh(Other)",
the other group of chinese users wouldn't be satisfied.

So, I propose to change the chinese setting into following situation (similar
with the old way):

In languagechooser, use

    zh    - ...(Choose this to proceed in Simplified Chinese)...
    zh    - ...(Choose this to proceed in Traditional Chinese)...

The first one use zh_CN translation, the second one use zh_TW (though there is
no traditional chinese translation yet -_- ). And whichever user selected,
countrychooser should always be triggered.

In countrychooser, the simplest way is that show all countries which use chinese
language.

         China
         Taiwan
         HongKong
         Other

The best way is that show the relevant countries accroding which kind of chinese
was choosen.

    1st, Simplified Chinese        2nd, Traditional Chinese
         China                          Taiwan
         Other                          HongKong
                                        Other

Then, debian-installer can show the correct language interface and get the right
locale.

Please give me your opinion. Thanks.


-- 
 Best Regards,
 Carlos