EUC-KP support (Re: [ko-KP] Mozilla Localization project)

mirangwt wrote:

> And We want to support [euc-KP] code of mozilla.

This is not a L10N issue, but an I18N issue (cross-posted to i18n group 
and followup-to set accordingly). Anyway, do you have any web
documents in EUC-KP? I've never seen one. If there isn't any (or are 
very few), there's no(very little) need to bloat Mozilla by adding yet 
another character encoding converter  although it's rather trivial to do 
so (except for one thing [1]) once you complete a conversion table 
between KPS 9566 and Unicode. The mapping between Korean syllables and 
Hanjas in KPS 9566 and those in Unicode is easy to generate, but it 
would be rather tedious to generate the mapping for the rest of 
characters in KPS 9566 ('symbol' characters).

You'd be much better off just going straight to UTF-8. South Korean web 
authors are still 'dwelling on' EUC-KR not because of its being better 
than UTF-8 but because of sheer 'inertia'. Not so much 'burdened' [2] 
with such a legacy and inertia as South Koreans, you (North Koreans) can 
just begin afresh with 'UTF-8'. Note also that virtually all recent 
Linux distributions come with UTF-8 locales *by default*.

In the past, the DPRK delegation to ISO/IEC JTC1/SC?/WG2 (where ISO 
10646 is dealt with) expressed concerns about, among other things,  the 
code point assignment of Korean syllables ('Hangul' in South Korea), 
which is ordered according to South Korean dictionary collation order. 
However, that should NOT be any concern at all. No serious program would 
sort per 'code point order'. Sort by 'code point order' NEVER works for 
any of languages. It does NOT work even for English, let alone Korean. 
What you need to do is to generate LC_COLLATE for glibc along with other 
parts of ko_KP.UTF-8 locale and submit it to the glibc maintainer.

I wrote the above (that may sound off-topic-ish here) to persuade you to 
give up EUC-KP and to go directly to UTF-8.



[1] There are four syllables encoded multiple times in KPS 9566.
[2] I don't think it's so much burden, but it's rather hard to make one 
a 'convert' for UTF-8. I've been 'campaigning' for all UTF-8 for the 
last 3 years, but it's moving slowly.