[tex-live] language, lang, la (was: testing TL'08)
arthur.reutenauer at normalesup.org
Sun Jun 15 20:49:17 CEST 2008
> We're talking about languages, so we should use the language codes
> (not country codes), i.e. ISO 639-3 (e.g. deu, slk, slv), or IETF
> language tags.
I strongly advocate for IETF tags (http://tools.ietf.org/rfc/bcp/bcp47.txt).
They are more precise than ISO 639 tags because you can add tags for
scripts (ISO 15924) as well as for countries (ISO 3166), as well as
other elements, if you need to. As far as the doc files are concerned,
we do actually need this for Chinese and Mongolian, because each of
these languages can be written in two different scripts (Traditional and
Simplified for Chinese -- ISO 15924 Hans and Hant, and Mongolian or
Cyrillic for Mongolian -- ISO 15924 Mong and Cyrl). The documents we
have here are in Simplified Chinese (zh-Hans) and Mongolian in Cyrillic
(mn-Cyrl). Until now, Chinese has been labelled as zh-CN in TeX Live --
I guess because it's the value of the POSIX locale -- but this does not
reflect what it really is (zh-CN would be Chinese from Mainland China,
which is roughly equivalent to Simplified Chinese, but it is still not
the same thing).
I find it is really important to choose a systematic scheme for
tagging languages. Mojca and I just reviewed all of TeX Live's pattern
files and converted them to UTF-8, and we also needed a naming scheme;
the IETF recommendation is the only one we found that could accomodate
all of our needs. It would really be a good thing if TeX Live could
have a clear naming policy for languages.
About czechslovak: I may be mistaken but I am under the impression
that texmf-doc/doc/czechslovak/texlive-cz/live.pdf is really a mixture
of Czech and Slovak -- most of it seems to Czech, but chapter 8 looks
like Slovak. Can anyone please clarify this? Zdeněk?
More information about the tex-live