W3C home > Mailing lists > Public > public-esw-thes@w3.org > October 2007

Re: [SKOS] Amsterdam topic "Concept semantics"

From: Antoine Isaac <aisaac@few.vu.nl>
Date: Wed, 03 Oct 2007 17:37:34 +0200
Message-ID: <4703B73E.3010708@few.vu.nl>
To: "Miles, AJ \(Alistair\)" <A.J.Miles@rl.ac.uk>
CC: SWD WG <public-swd-wg@w3.org>, public-esw-thes@w3.org

Hi Alistair,

I think your proposal makes sense. I would vote for the Primer, given 
that the subject is important. But perhaps we shall decide on that after 
the discussion.

Cheers,

Antoine

>> I agree that keeping the "domainless-ness" of SKOS 
>> labelling/documentation properties help solving a lot of 
>> problem (that's solution 4 in the wiki page).
>> One thing I would like to mention is that even with this 
>> option we should put something related on the subject of 
>> labelling/documentation of OWL classes in the current documentation.
>>     
>
> Yes, absolutely. I had imagined that we would write a document on different design patterns for using SKOS with OWL -- where using SKOS for labelling/documentation in OWL ontologies would be one of those design patterns. We could then include this document as either a section of the SKOS Primer, or publish as a separate Note on its own.
>
> What do you think?
>
> Cheers,
> Alistair.
>
>
>   
>> Actually it was in the draft I prepared, but the 
>> "wikification" caused the disappearing of the note :-(  
>> http://lists.w3.org/Archives/Public/public-swd-wg/2007Jun/0018.html
>> makes the point that applying labelling properties to a 
>> resource does not make it an instance of skos:Concept. 
>> However, there could be
>> confusion: all existing SKOS documentation (e.g . 
>> http://www.w3.org/2004/02/skos/core#prefLabel) refers to 
>> these properties as applied to instances of skos:Concept.
>>
>> Cheers,
>>
>> Antoine
>>
>> [1] http://www.w3.org/TR/swbp-skos-core-guide
>> [2] http://www.w3.org/TR/swbp-skos-core-spec
>>     
>>> +1
>>>
>>> Daniel
>>>
>>> At 08:20 AM 10/2/2007, Miles, AJ \(Alistair\) wrote:
>>>       
>>>> We have, as you mention, some use cases where people want 
>>>>         
>> to use only 
>>     
>>>> the labelling and documentation properties from SKOS, to add more 
>>>> human-readable content to their formal (OWL) ontologies.
>>>>
>>>> Currently, neither the SKOS labelling properties nor the SKOS 
>>>> documentation properties are declared with any domain. 
>>>>         
>> There are no 
>>     
>>>> dependencies between these properties and the skos:Concept class.
>>>>
>>>> I propose that we keep them like that. This would allow the SKOS 
>>>> labelling properties and the SKOS documentation properties to be 
>>>> treated as if they were standalone modules, and to be used 
>>>>         
>> anywhere 
>>     
>>>> in RDF or OWL, without having to worry about the semantics of the 
>>>> skos:Concept class.
>>>>
>>>> (Maybe we could give this a name, as a design pattern for 
>>>>         
>> using SKOS 
>>     
>>>> and OWL together -- i.e. "OWL + SKOS labelling and documentation 
>>>> only" or something like that?)
>>>>
>>>> Anyway, if we keep them like that, then we don't need to 
>>>>         
>> consider any 
>>     
>>>> of the SKOS labelling or documentation properties in our 
>>>>         
>> discussion 
>>     
>>>> of the semantics of skos:Concept. We can just focus on the 
>>>>         
>> semantics 
>>     
>>>> of skos:Concept, and design patterns for using 
>>>>         
>> skos:Concept with OWL 
>>     
>>>> classes, properties and individuals. This may simplify some of the 
>>>> options at [1].
>>>>
>>>> Cheers,
>>>>
>>>> Alistair.
>>>>         
>>>
>>>       
>>
>>     
>
>   
Received on Wednesday, 3 October 2007 16:41:40 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Monday, 7 December 2009 10:38:58 GMT