Hey

Our language system has been around for a while and with the many new targets
and new features added since its creation, we can now see huge cracks in the
way it functions.

I say it is about time we upgrade the concept and file formats to better deal
with these major features:

    A) Possibly different original strings depending on target player

    B) Possibly different translations depending on target player

    C) Possibly different voice strings depending on target player (and of
       course no voice at all for a string)

    D) No more having strings for other players around (like how Archos
       Recorders currently load Archos Player strings and vice versa)

I'm personally quite happy with how uplang works, as when run on an existing
translation it properly identifies phrases that the translator needs to take
care of. Therefore, I want to move to a format where we can keep this feature.

No matter what new format we come up with, we must make a converter from the
old format and we should maintain the existing generated source style/ABI.

Suggestion: (format slightly inspired by apache's config file syntax)

<phrase>
   id: LANG_ID_NAME
   desc: english descripton of where or how the phrase is used

   <source>
     *: "generic phrase"
     recorder: "phrase adjusted to the recorder"
     h100: "phrase adjuted to h100"
   </source>

   <dest>
     *: "generic translation"
     player: "translation adjusted for the player"
     h300: NONE # blanked for this target
   </dest>

   <voice>
     *: "generic voice string for this string"
     ondiosp: "voice adjusted for ths ondiosp"
   </voice>

</phrase>

I'm volounteering to write the scripts and stuff, I just want us to come up
with a fine file format first.

Of course, I don't think we should force this new concept in before the next
release. This is meant for later.

-- 
  Daniel Stenberg -- http://www.rockbox.org/ -- http://daniel.haxx.se/
_______________________________________________
http://cool.haxx.se/mailman/listinfo/rockbox