On Mon, 26 Nov 2001, Richard, Francois M wrote: > Reading http://oss.software.ibm.com/developerworks/opensource/locale/, my > understanding is that "Universal Locales for Linux" added a lot of utf-8 > based Locales to the standard set of Linux Locales. Theses Locales also > extend the POSIX Locale model by adding LC_NAME, LC_ADDRESS, LC_PAPER, > LC_MEASUREMENT and LC_TELEPHONE.
All these new LC_*'s are defined in ISO 14652 (specification methods for cultural conventions). See <http://anubis.dkuug.dk/jtc1/sc22/wg20> (ISO/IEC JTC1/SC22/WG20 : internationalization working group). BTW, locales included in glibc also 'support' them. So far I have found a couple of applications using LC_PAPER (pstools and apsfilter?). Other than that, I haven't found any program to make use of new LC_*'s. > Is this package standard on Linux distributions (like redhat)? I don't think so. At least RH 7.x and Mandrake 8.x shipped C lib. locales from glibc distribution. > How do I know > if it has been installed on a Linux machine? Is this synchronization between > Locale models (Linux and ICU) well accepted and used? I guess synchronization is a nice thing to have. I guess some features offered by ICU locale model are not available C lib. locale model, though. > What are its > advantages since if I am relying on ICU, I do not use Linux Locales (no > SetLocale() in my programs), but the ones provided by ICU? If you don't use C lib. locales and exclusively rely on ICU locales, I don't think you don't have to worry about which 'version' of C lib. locales are installed, those from glibc distribution or those from IBM. Hmm, this should be obvious... Did you meant something else? One area in which ICU locales are very good is collation (1. ISO 14651 : international string ordering also available at ISO/IEC JTC1/SC22/WG20 web page 2. Unicode TR #10: Unicode collation algorithm) Jungshik Shin -- Linux-UTF8: i18n of Linux on all levels Archive: http://mail.nl.linux.org/linux-utf8/
