Re: [issue 3] What ontology should describe the metadata values (entity types)

Hi Maxime, all,
Yes, the general principle is that a data category should be defined in 
a schema/language agnostic way, which for a pointer to an external 
concept or term means it must be agnostic to both the schema/language of 
the document hosting the pointer (though we are currently restricting 
this to HTML5 and XML) and to the schema/language of the document (or 
database?) hosting what is being pointed at.

One option is just use IRIs using a single pointer attribute That 
happily addresses things we want to reference in the LOD and semantic 
web world, and quite a lot of other language resources are moving in 
this direction, e.g. NIG, Olif etc.

However, this leave us with the problems of addressing other important 
file formats such as TBX, TMX and XLIFF.  Does anyone have experience of 
best practice in using external references to elements in such formats, 
e.g. using Xpointer, URI with specific conventions for resolving element 
ID?

Would we need to identify the file type in the data category if 
different mechanisms are needed for different formats (see earlier 
posting on lack of mime type for these files)?

cheers,
Dave

On 27/04/2012 14:58, Maxime Lefrançois wrote:
> Hi,
>
> Considering that obviously no vocabulary will be sufficient for all 
> usage (open world assumption), I say:
>  - describe how a specific schema can be used in a standard manner, 
> without restricting its choice in any ways.
> Thus, for named entities and terms, an organization can indifferently 
> use the linked open data cloud, or one of its specific ontology or 
> schema or vocabulary or tbx or tmx or xliff or whatever...
>
> Best,
> Maxime Lefrançois
> Ph.D. Student, INRIA - WIMMICS Team
> http://maxime-lefrancois.info <http://maxime-lefrancois.info/>
> @Max_Lefrancois <http://twitter.com/Max_Lefrancois>
>