Re: feedback sought: using owl:imports

Hi Carlos,

A few thoughts below:


Carlos A Velasco wrote:
> Hi all,
> 
> Here is my view.
> 
> Shadi Abou-Zahra wrote:
>> Dear group,
>>
>> Ref: <http://www.w3.org/TR/owl-ref/#imports-def>
>>
>> There is a request to use the owl:imports statement in the RDF files of
>> the different EARL specifications. According to the OWL specification,
>> the owl:imports statement is defined as follows:
>>
>>  - "An owl:imports statement references another OWL ontology containing
>> definitions, whose meaning is considered to be part of the meaning of
>> the importing ontology."
> 
> I read this differently than Shadi. I do not think the import statement
> implies that you must adopt *all definitions*, but only those that you
> refer to. Thus ...

The entire section defining owl:imports does not mention any of what you 
say but rather talks about entire ontologies importing others. It also 
talks about tools loading these imported ontologies but does not in any 
way mention individual statements from ontologies.

Are you aware of how OWL tools use the owl:imports statement? Also, what 
is actually the benefit of using owl:imports? What are the downsides?


>> # Scenario A:
>>  - do not adopt owl:imports (currently the EARL 1.0 Schema RDF is using
>> rdfs:seeAlso which expresses a relationship between two ontologies)
> 
> -1
> 
>> # Scenario B:
>>  - EARL 1.0 Schema imports HTTP-in-RDF, Content-in-RDF, and Pointers-in-RDF
> 
> -1
> IMHO we miss important bits, like DCT and FOAF
> 
>> # Scenario C:
>>  - EARL 1.0 Schema adopts HTTP-in-RDF, and Pointers-in-RDF
>>  - HTTP-in-RDF imports Content-in-RDF
> 
> -1
> 
>> # Scenario D:
>>  - EARL 1.0 Schema imports HTTP-in-RDF, Content-in-RDF, Pointers-in-RDF,
>> DC, and FOAF
> 
> -1
> IMHO HTTP-in-RDF, Content-in-RDF, Pointers-in-RDF must also import
> 
>> # Scenario E:
>>  - EARL 1.0 Schema imports HTTP-in-RDF, Content-in-RDF, Pointers-in-RDF,
>> DC, and FOAF
>>  - HTTP-in-RDF imports Content-in-RDF, DC, and FOAF
>>  - Content-in-RDF imports DC and FOAF
>>  - Pointers-in-RDF imports DC and FOAF
> 
> +1
> *BUT* - I prefer dct prefix for DC Terms, to avoid confusion with legacy
> apps
> *BUT* - Content-in-RDF imports only DCT
> *BUT* - HTTP-in-RDF imports only Content-in-RDF, DCT
> *BUT* - Pointers-in-RDF imports only Content-in-RDF (I did not find any
> ref to DCT)

This would make HTTP-in-RDF, Content-in-RDF, and Pointers-in-RDF to be 
*equivalent* ontologies. Is that the meaning we actually want to have?


Regards,
   Shadi


>> # Scenario F:
>>  - EARL 1.0 Schema adopts HTTP-in-RDF, Pointers-in-RDF, DC, and FOAF
>>  - HTTP-in-RDF imports Content-in-RDF, DC, and FOAF
>>  - Content-in-RDF imports DC and FOAF
>>  - Pointers-in-RDF imports DC and FOAF
> 
> -1
> 
>> Note 1: there was a remark on one of the calls about importing back, for
>> instance that HTTP-in-RDF should import EARL 1.0 Schema. However, the
>> OWL specification specifically points out that this would mean that
>> HTTP-in-RDF and EARL 1.0 Schema are *equivalent*.
> 
> -1
> 
>> Note 2: the OWL specification points out that an imported ontology is
>> part of the *meaning* of the importing ontology, which would make DC and
>> FOAF directly part of EARL in scenarios D-F.
> 
> Like I said, I read this differently. In a sense, we need the import
> statement to refer to the definitions of what we use from DCT and FOAF.
> 
> regards,
> carlos
> 

-- 
Shadi Abou-Zahra - http://www.w3.org/People/shadi/ |
   WAI International Program Office Activity Lead   |
  W3C Evaluation & Repair Tools Working Group Chair |

Received on Monday, 29 June 2009 22:17:07 UTC