Alaa Abd El Fattah wrote: > On Tue, 09 Mar 2010 13:56:14 +0200 > Ihar Hrachyshka <[email protected]> wrote: > >> On Wed, 2010-03-10 at 00:39 +1300, Amos Jeffries wrote: >>> Christian PERRIER wrote: >>>> Quoting Amos Jeffries ([email protected]): >>>> >>>>> Problem 1) Alphabets versus Languages >>>>> I've hit it with Serbian. They use two different alphabets >>>>> Latin and Cyrillic. But only one language. >>>>> Distinguished by two codes sr-Latn and sr-Cyrl. The same issue >>>>> occurs in Chinese Hans/Hant/Ming/* and has been hacked around >>>>> previously by appending the specific ISO-3166 country code where >>>>> its most frequently needed. >>>>> >>>>> What I'm hoping for is to use the ISO-3066 alphabet codes as >>>>> part of the language tag somewhere. >>>> >>>> This is indeed the first time I hear about ISO-3066. >>>> >>>> As one of the iso-codes maintainers, I know about ISO-15924, >>>> which is meant to be a standard for script names. We include it >>>> in the package since October 2007. Reference is >>>> http://unicode.org/iso15924/ >>> Ah thanks. Good to know. >>> >>>> Example entry in the XML file we provide: >>>> >>>> <iso_15924_entry >>>> alpha_4_code="Cyrl" >>>> numeric_code="220" >>>> name="Cyrillic" /> >>>> <iso_15924_entry >>>> alpha_4_code="Cyrs" >>>> numeric_code="221" >>>> name="Cyrillic (Old Church Slavonic variant)" /> >>>> .../... >>>> <iso_15924_entry >>>> alpha_4_code="Latn" >>>> numeric_code="215" >>>> name="Latin" /> >>>> >>>> >>>> These examples use your own example. Note that the alpha4 code is >>>> indeed the same. >>>> >>>> I'd say that ISO-15924 seems to be an evolution of 3066 or >>>> something like this. >>> I guess so. I only found the ISO-3066 code this week in some fairly >>> old university language papers about Serbian/Croatian alphabet >>> splits. >>> >>>> WRT your general message, I agree that using ISO 15924 codes in >>>> locale names would be a great progress over the current hacks >>>> implemented in various ways (zh_CN vs. zh_TW as a hack between >>>> Simplified and Traditional Chinese....or "Hans" vs. "Hant", or >>>> variants for Serbian, or probably others I don't know about). >>>> >>> So far I know of Chinese and Serbian for certain, with hints >>> indicating Azerbaijan and Croatian will need it in future as well. >> ...and Belarusian Latin is assigned to "b...@latin" in glibc (IIRC >> Serbian uses '@Latn' tag for the same thing). Actually, these locale >> 'variants' don't have good support in different l10n software (f.e. >> Rosetta doesn't know about their existance at all). > > Poolte uses glibc locale's and supports codes like b...@latin, they're > inconsistently used for other types of variations like c...@valencia but > the good news is they work fine with our tools > (check http://pootle.locamotion.org/c...@valencia/ for example). > > I'm not sure I understood the issues Amos is facing, how much of it is > solved by using s...@latin? >
My problem #1 can be resolved completely by s...@latin. Thanks for pointing it out. I had seen c...@valencia without really understanding what that was about, it slipped my mind. But ... where do I find a reliable index of these @... codes? searching online for stuff with '@' in it seems to be one of the difficult tasks, and even "@valencia" did not lead anywhere useful. FYI: The web standard my raw .po files have to use in VCS uses '-' instead of '@' and the ISO-15924 codes instead of "valencia" or "latin" glibc codes. Otherwise identical in meaning. My problem #2 is partly about needing to store man page translations (with system Locales) and these web-format translations side by side for each language. ie s...@latin/sr-Latn.po, s...@latin/sr_SP.po s...@latin/sr_SB.po s...@cyrillic/sr-Cyrl.po, s...@cyrillic/sr_SP.po s...@cyrillic/sr_SB.po Or do I need the .pot name in the .po filename like Rosetta appear to use? Amos Squid Project ------------------------------------------------------------------------------ Download Intel® Parallel Studio Eval Try the new software tools for yourself. Speed compiling, find bugs proactively, and fine-tune applications for parallel performance. See why Intel Parallel Studio got high marks during beta. http://p.sf.net/sfu/intel-sw-dev _______________________________________________ Translate-pootle mailing list [email protected] https://lists.sourceforge.net/lists/listinfo/translate-pootle
