Re: translations

Miles, AJ (Alistair) wrote:

>Hi Tom,
>
>  
>
>>That is one way to do it, but it ultimately puts a significant
>>maintenance burden on the person (i.e., you) who is collecting
>>and aggregating those translations into one big schema.  
>>
>>As an alternative -- in principle, at any rate -- translators
>>might maintain the translations in RDF on their own servers
>>and you would merely point to them and merge their contents
>>into a multi-lingual RDF description automatically.
>>    
>>
>
>I think we probably need all RDF descriptions of SKOS Core to be under the same configuration management, so it may be a problem for SKOS Core to import files from other servers.
>
>What if we have on the W3C server e.g.
>
>core.rdf
>core/annotations/en.rdf
>core/annotations/fr.rdf
>
>  
>
yup (though i'm also happy having translations managed elsewhere,
once we're more stable...).

>... and so on, with core.rdf containing statements like ...
>
><http://www.w3.org/2004/02/skos/core> owl:imports <http://www.w3.org/2004/02/skos/core/annotations/fr> .
>  
>
I'm not sure owl:imports is very fashionable any more. I used to use it
in FOAF but
someone told me it wasn't good for DL tools, or something. Maybe I was
importing
the owl.owl or the RDF namespace? I forget, sorry for the murky memory.

I'm happy with 'seeAlso' anyway...


>... for each language resource, and then e.g. core/annotations/fr.rdf containing only statements involving predicates with literal range (i.e. rdfs:label, rdfs:comment, skos:definition etc.)?
>
>Or do we want http://www.w3.org/2004/02/skos/core to serve a content-negotiated representation with only the annotations in the requested language? 
>
>  
>
Better not to hide the translations, I think. They're for machines
rather than
people anyhow...

Dan