Hi Folks...

Yoshiki proposed the idea of creating a chain of dictionaries where the
deepest place is a type of application-specific translation table.  As I
see the problem is something like that: The package owner can provide a
set of phrases to translate and different translators will produce
translation-dictionaries (one per language).  In this case we need a type
of double-entry lookup (the application and the language), in this case
the methods in String will be something like that.

   String>>translated
      "answer the receiver translated to the current language using the
topmost phrases set"

   String>>translatedTo: aLanguage
      "answer the receiver translated to aLanguage using the topmost
phrases set"

   String>>translatedFor: translationSet
      "answer the receiver translated to the current language using the
translationSet"

   String>>translatedTo: aLanguage for: translationSet
      "answer the receiver translated to aLanguage using the
translationSet"

On the other side the translation tool has to show all the translation
sets and offer a way to look for phrases all over the sets, etc.

I don't think this extra complexity pay too many benefits to the option of
having only one translation-set. Yoshiki: What I missed?

Cheers,

Diego Gomez Deck
http://www.small-land.org