> > I am wondering, if it's effective to use text arrays to store
> > multilanguage information.
We solved it like this:
http://savannah.gnu.org/cgi-bin/viewcvs/*checkout*/gnumed/gnumed/gnumed/server/sql/gmI18N.sql?rev=1.15
- i18n_curr_lang holds the (ISO) language string per user
- i18n_keys holds the strings to be translated
- i18n_translations holds the translations
- function i18n() inserts a text value into i18n_keys thus
marking it for translation, this we use during insertion of
data in the "primary language" (think gettext, think English)
- function _() returns a translated text value (or it's
primary language equivalent), we use that in queries and view
definitions
- v_missing_translations is a view of, well, ...
Then here:
http://savannah.gnu.org/cgi-bin/viewcvs/*checkout*/gnumed/gnumed/gnumed/server/locale/dump-missing-db_translations.py?rev=1.3
we've got a Python tool that will connect to your database and
create schema files ready for re-insertion via psql after
adding the missing translations. Kind of like a gettext po
file.
Karsten
--
GPG key ID E4071346 @ wwwkeys.pgp.net
E167 67FD A291 2BEA 73BD 4537 78B9 A9F9 E407 1346