Re: SKOS-ISSUE:ConceptSchemesContainment

Hello Sean and Tom,

+1 for Tom's comments.

Actually I think this issue applies further than the case of a 
relationship between two concepts. It could well apply to any triple a 
concept scheme asserts. What if two concept schemes use a same concept, 
butwith the first one having it as as a accepted since 1995, while the 
second only includes it since 1999, or if they have different status 
(stable versus unstable), and if the concept scheme managers want to 
express these things in their SKOS representations using SKOS notes that 
can point at literals? My examples may sound tricky, but I think they 
mirror some real concern and bring a motivation for going further than 
just concept relationships.

Antoine

>On Tue, Jan 09, 2007 at 04:04:42PM +0000, Sean Bechhofer wrote:
>  
>
>>TITLE:Relationships in Concept Schemes
>>DESCRIPTION: SKOS provides a notion of Concept Schemes. RDF's triple  
>>syntax makes it impossible to represent associations between concept  
>>schemes and particular relationships (e.g. a BT relationship) without  
>>resorting to reification. A principled approach to representing this  
>>containment would desirable.
>>RAISED BY: Sean Bechhofer, 08/01/07
>>REFERENCE: http://lists.w3.org/Archives/Public/public-swd-wg/2006Nov/ 
>>0078.html
>>    
>>
>
>Sean,
>
>The wording ("represent associations between concept schemes
>and particular relationships") seems confusingly terse if
>what you want to represent is "the fact that the semantic
>relationships between concepts occur within a particular
>scheme" [1]. 
>
>For example, I think you mean not just a "particular [type of]
>relationship, such as a BT relationship", but "a particular
>BT relationship between two concepts".  Also, "associations
>between concept schemes" per se (e.g., "this concept scheme is
>associated with that concept scheme") are not the issue here,
>but the sentence could be read this way.
>
>Maybe something like: "RDF's triple syntax makes it impossible
>to associate a particular relationship between two concepts
>with a concept scheme within which that relationship is
>contained"?
>
>A minor point, but I suggest that instead of the date format
>"08/01/07" (August 1 in US), we use "2007-01-08".
>
>Tom
>
>[1] http://lists.w3.org/Archives/Public/public-swd-wg/2006Nov/0078.html
>
>  
>

Received on Wednesday, 10 January 2007 10:42:05 UTC