Identify "Corpus" by IETF language tag rather than just language

[duramato]

My suggestion is to Identify the "Corpus" by IETF language tag rather than just language. What are the benefits? it would allow for different dialects to be made independent as i might want to just "teach" just one of them.

Why i'm bringging this idea up?

For example the corpus for the Portuguese language taking a look at it seems to have mostly Portuguese Brazilian (pt-BR) strings and some, here and there, in the Portuguese (pt or pt-PT). Using said corpus makes the bot a bit of biliangual freak, i'f im allowed to call it that :P

Same goes for the english corpus which, i'd say (not totally sure, but from some expressions), is English (en or en-GB) with United States English (en-US) in it.

Chinese is another language with ALOT of dialects.. (but this one is an unkown to me as i have zero knowledge in the language)
The story goes on..

[gunthercox]

This is a good suggestion. Making changes to use IETF language tags sounds like it would really benefit ChatterBot.