[rdfweb-dev] languages

[Peter Saint-Andre]

On Wed, Jul 30, 2003 at 04:04:20PM +0100, Ian Davis wrote:

> A person can speak, read, write or understand one or more languages.
> Each language can independently be spoken, read, written or understood
> with varying proficiencies. A person may rank their preferred
> languages for each mode of communication.
> 
> Maybe proficiencies aren't absolute but are relative.
> 
> Here's one suggestion for implementing this:
> 
> <lang:speaks>
>   <lang:LanguageProficiency rdf:ID="spokenEnglish">
>     <lang:iso639Code>en</lang:iso639Code>
>   </lang:LanguageProficiency>
> </lang:speaks>
> 
> <lang:speaks>
>   <lang:LanguageProficiency rdf:ID="spokenFrench">
>     <lang:iso639Code>fr</lang:iso639Code>
>     <lang:wouldPrefer rdf:resource="#spokenEnglish" />
>   </lang:LanguageProficiency>
> </lang:speaks>
> 
> <lang:speaks>
>   <lang:LanguageProficiency rdf:ID="spokenGerman">
>     <lang:iso639Code>de</lang:iso639Code>
>     <lang:wouldPrefer rdf:resource="#spokenEnglish" />
>     <lang:wouldPrefer rdf:resource="#spokenFrench" />
>   </lang:LanguageProficiency>
> </lang:speaks>
> 
> An agent trying to determine the best common language picks one that
> both parties understand at random. It checks the wouldPrefer property
> to see if there's a better language that they both speak. Continue
> until no more wouldPrefer properties.
> 
> Just more ingredients for the pot...

And tasty ingredients at that. :)

I like the idea of a cardinal ranking of language preferences. I'm a bit
confused by the LanguageProficiency class, since it does not seem to
capture my proficiency level (perhaps that really isn't necessary, but I
expected it to be there given the class name).

One nit about Leandro's schema is that it explicitly forbids the use of
ISO 639-2 identifiers, whereas I'd like to include those since I read
Ancient Greek and there is no 639-1 identifier for that language.

Peter

-- 
Peter Saint-Andre
Jabber Software Foundation
http://www.jabber.org/people/stpeter.php