|
|
Posted by Katana on 10/01/09 11:23
> Would you recommend solutions which use a marker inside the sources
> to mark strings for translation, leaving the original strings inside
> the sources, like gettext does ('t("no messages")'), or prefer a
> solution which only refers to abstract hierarchical IDs inside the
> sources, for example like 't("msg.list.nomsgs")'?
If you go for the DB driven solution, you have to implement caching
using flat files, definitely.
As far as I'm concerned, I think that using a DB here doesn't make much
sense... you would be reinventing the wheel. Whatever you do, you have
to identify the strings, so you need to run a script or something in
order to index the content...
I would definitely suggest the good old gettext approach :)
Bertrand
Navigation:
[Reply to this message]
|