comment/query re: Simple part-whole relations in OWL Ontologies

Previous Topic Next Topic
 
classic Classic list List threaded Threaded
3 messages Options
Reply | Threaded
Open this post in threaded view
|

comment/query re: Simple part-whole relations in OWL Ontologies

Danny Ayers

My query is primarily about the current status of this doc. I assumed
there would be a Note somewhere on part-whole relationships, Google
led me to this W3C Editor's Draft 11 Aug 2005 [1]. The status given on
the SWBPD page [2] is "Reviewed and ready for note status - requires
html cleanup.".  Aside from the time lag, there is at least one issue
that's slipped through review, see below.

Since SKOS has come along it seems likely that may offer another
approach to the part-whole problem (anyone have any pointers on
this?).

The draft includes the following:
[[
The code for classes and properties in RDF Schema

partOf
  a rdf:Property .

partOf_directly
  a rdf:Property ;
  rdfs:subPropertyOf partOf .

Car
  a rdfs:Class .

Engine
  a rdfs:Class .

Headlight
  a rdfs:Class .

Wheel
  a rdfs:Class .

As shown, there is not much one can express in RDFS, and as a result
it will not be used for the remainder of this note.
]]

Aside from that final sentence having an unnecessarily dismissive
tone, it's plainly inaccurate - statements about the classes and/or
individuals in those classes can be made using properties, as shown in
the OWL-oriented examples that follow, e.g.

[[
engine123
  a Engine ;
  part:partOf_directly car123 .
]]

I realise the draft is aimed at describing the problem in the context
of OWL ontologies, but my own requirement which led me to search for
this is to initially express part-whole relationships in RDF(S), so I
can use SPARQL one the data. I'm not sure at this point in time if
I'll want DL inference further down the road, or whether there are
OWL-based tools that could potentially be used with the data. For this
general reason I personally believe that it's good practice when using
RDF(S) to remain consistent with OWL DL unless there are strong
reasons not to, but the draft doesn't address this option - it takes a
stance of "OWL or nothing".

Cheers,
Danny.

[1] http://www.w3.org/2001/sw/BestPractices/OEP/SimplePartWhole/
[2] http://www.w3.org/2001/sw/BestPractices/OEP/

--

http://dannyayers.com

Reply | Threaded
Open this post in threaded view
|

Re: comment/query re: Simple part-whole relations in OWL Ontologies

David Wood-3

Hi Danny,

The Semantic Web Best Practices and Deployment Working Group was  
closed on 29 September 2006.  Its successor is the Semantic Web  
Deployment Working Group.  Only some of the activities of the old  
group were carried forward to the new one.  The part-whole work was  
unfortunately not one of them.  That is, there is no current part-
whole work at the W3C to the best of my knowledge.

I am not certain whether anyone has fully formulated a part-whole  
approach in SKOS.  I presume you have been monitoring the public-esw-
[hidden email] mailing list?  If not, that would be the place to ask.

Regards,
Dave


On 2 Apr2007, at 06:41, Danny Ayers wrote:

>
> My query is primarily about the current status of this doc. I assumed
> there would be a Note somewhere on part-whole relationships, Google
> led me to this W3C Editor's Draft 11 Aug 2005 [1]. The status given on
> the SWBPD page [2] is "Reviewed and ready for note status - requires
> html cleanup.".  Aside from the time lag, there is at least one issue
> that's slipped through review, see below.
>
> Since SKOS has come along it seems likely that may offer another
> approach to the part-whole problem (anyone have any pointers on
> this?).
>
> The draft includes the following:
> [[
> The code for classes and properties in RDF Schema
>
> partOf
>  a rdf:Property .
>
> partOf_directly
>  a rdf:Property ;
>  rdfs:subPropertyOf partOf .
>
> Car
>  a rdfs:Class .
>
> Engine
>  a rdfs:Class .
>
> Headlight
>  a rdfs:Class .
>
> Wheel
>  a rdfs:Class .
>
> As shown, there is not much one can express in RDFS, and as a result
> it will not be used for the remainder of this note.
> ]]
>
> Aside from that final sentence having an unnecessarily dismissive
> tone, it's plainly inaccurate - statements about the classes and/or
> individuals in those classes can be made using properties, as shown in
> the OWL-oriented examples that follow, e.g.
>
> [[
> engine123
>  a Engine ;
>  part:partOf_directly car123 .
> ]]
>
> I realise the draft is aimed at describing the problem in the context
> of OWL ontologies, but my own requirement which led me to search for
> this is to initially express part-whole relationships in RDF(S), so I
> can use SPARQL one the data. I'm not sure at this point in time if
> I'll want DL inference further down the road, or whether there are
> OWL-based tools that could potentially be used with the data. For this
> general reason I personally believe that it's good practice when using
> RDF(S) to remain consistent with OWL DL unless there are strong
> reasons not to, but the draft doesn't address this option - it takes a
> stance of "OWL or nothing".
>
> Cheers,
> Danny.
>
> [1] http://www.w3.org/2001/sw/BestPractices/OEP/SimplePartWhole/
> [2] http://www.w3.org/2001/sw/BestPractices/OEP/
>
> --
>
> http://dannyayers.com
>


Reply | Threaded
Open this post in threaded view
|

Re: comment/query re: Simple part-whole relations in OWL Ontologies

Danny Ayers

On 02/04/07, David Wood <[hidden email]> wrote:
> Hi Danny,
>
> The Semantic Web Best Practices and Deployment Working Group was
> closed on 29 September 2006.  Its successor is the Semantic Web
> Deployment Working Group.  Only some of the activities of the old
> group were carried forward to the new one.  The part-whole work was
> unfortunately not one of them.  That is, there is no current part-
> whole work at the W3C to the best of my knowledge.

Ah, of course, that explains it. Thanks David.

> I am not certain whether anyone has fully formulated a part-whole
> approach in SKOS.  I presume you have been monitoring the public-esw-
> [hidden email] mailing list?

Heh, I'm subscribed but not exactly monitoring...

> If not, that would be the place to ask.

Got it.

Thanks again,
Danny.


--

http://dannyayers.com