W3C home > Mailing lists > Public > public-swd-wg@w3.org > March 2009

Re: [SKOS] SKOS ontology sanity-check? An attempt at creating OWL2 axioms for SKOS

From: Antoine Isaac <aisaac@few.vu.nl>
Date: Sun, 22 Mar 2009 17:17:42 +0100
Message-ID: <49C664A6.9000302@few.vu.nl>
To: Ivan Herman <ivan@w3.org>
CC: SWD WG <public-swd-wg@w3.org>
Hi Ivan,

Thank you very much for the feedback! That's a great help.
I note the RDF examples of [1], but still miss the RDF/XML version ;-)

Cheers,

Antoine




> Antoine,
> 
> a heroic effort with a moving target:-)
> 
> Just a few pointers that may help:
> 
> - the editor's draft of the Syntax document for OWL 2[1] includes all
> the examples in RDF, too. There is a set of buttons at the beginning of
> the document where you can select which syntax you want to see for the
> examples. Maybe that helps...
> 
> - it is indeed true that the core functional syntax for OWL 2 does not
> define the usage of property disjointness for annotation properties.
> However, in the RDF compatible semantics of OWL 2 (essentially, OWL 2
> Full), propertyDisjointWith is defined on _any_ property, ie, it
> provides the necessary semantics for disjoint annotation properties, too.
> 
> Disjointness of annotation properties as a requirement never came to the
> OWL 2 group, to be honest...
> 
> - as an answer to one of the LC comments, the RDF encoding of property
> chains have changed. In the final version of OWL 2 the encoding of your
> axiom should be (in turtle):
> 
> skos:prefLabel
>    owl:propertyChainAxiom (skos:prefLabel skos:literalForm).
> 
> 
> Cheers
> 
> Ivan
> 
> 
> [1] http://www.w3.org/2007/OWL/wiki/Syntax
> [2] http://www.w3.org/2007/OWL/wiki/RDF-Based_Semantics
> 
> Antoine Isaac wrote:
>> Hi,
>>
>> I have given a try at creating OWL 2 axioms for representating our
>> non-formally stated axioms in skos.rdf and [1] for skos-xl.rdf [2], see
>> below.
>>
>> I found it in fact really more difficult than planned, and the lack of
>> RDF/XML examples in OWL2 is really an obstacle.
>>
>> Anyway, it seems a good idea to already be aware right now of the issues
>> that we could have later...
>>
>> I hope this will help,
>>
>> Antoine
>>
>> ==================
>>
>>    S12.  The rdfs:range of each of skos:prefLabel, skos:altLabel and
>> skos:hiddenLabel is the class of RDF plain literals.
>>
>> At first sight this axiom could represented using rdf:Text. [3] reads:
>>> The datatype identified by the URI
>>> http://www.w3.org/1999/02/22-rdf-syntax-ns#text (abbreviated rdf:text)
>>> allows for the representation of internationalized text strings. In
>>> addition to the RIF and OWL specifications, this datatype is expected
>>> to supersede RDF's plain literals with language tags
>> However, I would not recommend to use it until the "expectations" have
>> been met, and RDF plain literals indeed been superseded.
>>
>> ==================
>>
>>    S13. skos:prefLabel, skos:altLabel and skos:hiddenLabel are pairwise
>> disjoint
>>
>> Here is a first try at an RDF/XML representation of this axiom:
>>
>>    <owl:ObjectProperty rdf:about="#prefLabel">
>>        <owl:propertyDisjointWith rdf:resource="#altLabel"/>
>>        <owl:propertyDisjointWith rdf:resource="#hiddenLabel"/>
>>    </owl:ObjectProperty>
>>    <owl:ObjectProperty rdf:about="#altLabel">
>>        <owl:propertyDisjointWith rdf:resource="#hiddenLabel"/>
>>    </owl:ObjectProperty>
>>
>>
>> However I'm not sure this would be ok, as [4] hints at applying it
>> either to object properties or datatype properties, but says nothing on
>> annotation properties...
>>
>> Indeed I tried to create the axiom using Protégé 4. It assumed that
>> prefLabel, altLabel and hiddenLabel were instances of DatatypeProperty,
>> and represented my axioms as:
>>        <owl:disjointObjectProperties rdf:resource="#hiddenLabel"/>
>> But apparently this owl:disjointObjectProperties construct does not
>> exist in the RDF mapping at [5]!
>>
>> Note that OWL 2 also allows for group declaration of disjoint
>> properties. But given that we only have three disjoint properties, and
>> that group disjointness seems cumbersome to represent in RDF syntax (it
>> requires using owl:AllDisjointProperties, and an owl:members which
>> intuitively conflicts with rdf:member the same way as our own
>> skos:member does :-/ ), I'd recommend to keep to the simple construct.
>>
>>
>> ==================
>>
>>  S14. A resource has no more than one value of skos:prefLabel per
>> language tag.
>>
>> This cannot be expressed.
>>
>> =================
>>
>>  S27. skos:related is disjoint with skos:broaderTransitive
>>
>> With the same caveats as for S13 (my Protégé had used
>> owl:disjointObjectProperties), we can represent this axiom by:
>>
>>    <owl:ObjectProperty rdf:about="#related">
>>        <owl:propertyDisjointWith rdf:resource="#broaderTransitive"/>
>>    </owl:ObjectProperty>
>>
>> =================
>>
>>    S36. For any resource, every item in the list given as the value of the
>>      skos:memberList property is also a value of the skos:member property
>>
>> This axiom cannot be represented, unless we first introduce an auxiliary
>> propery that allows a direct access to the object of the rdf:rest
>> statement in the definition of the member list.
>>
>> =================
>>
>>    S46. skos:exactMatch is disjoint with each of the properties
>> skos:broadMatch and skos:relatedMatch.
>>
>> With the same caveats as for S13, we can represent this axiom by:
>>
>>    <owl:ObjectProperty rdf:about="#exactMatch">
>>        <owl:propertyDisjointWith rdf:resource="#broadMatch"/>
>>        <owl:propertyDisjointWith rdf:resource="#relatedMatch"/>
>>    </owl:ObjectProperty>
>>
>>
>> ============ SKOS-XL
>>
>>    S51. The range of skosxl:literalForm is the class of RDF plain literals.
>>
>> Cf S12.
>>
>> =================
>>
>>    S55. If C skosxl:prefLabel L and L skosxl:literalForm V, then X
>> skos:prefLabel V.
>>
>>    <rdf:Description>
>>        <rdfs:subPropertyOf
>> rdf:resource="http://www.w3.org/2004/02/skos/core#prefLabel"/>
>>        <owl:propertyChain rdf:parseType="Collection">
>>            <rdf:Description
>> rdf:about="http://www.w3.org/2008/05/skos-xl#prefLabel"/>
>>            <rdf:Description
>> rdf:about="http://www.w3.org/2008/05/skos-xl#literalForm"/>
>>        </owl:propertyChain>
>>    </rdf:Description>
>>
>>
>>
>> =================
>>
>>    S56. If C skosxl:altLabel L and L skosxl:literalForm V, then X
>> skos:altLabel V.
>>
>>    <rdf:Description>
>>        <rdfs:subPropertyOf
>> rdf:resource="http://www.w3.org/2004/02/skos/core#altLabel"/>
>>        <owl:propertyChain rdf:parseType="Collection">
>>            <rdf:Description
>> rdf:about="http://www.w3.org/2008/05/skos-xl#altLabel"/>
>>            <rdf:Description
>> rdf:about="http://www.w3.org/2008/05/skos-xl#literalForm"/>
>>        </owl:propertyChain>
>>    </rdf:Description>
>>
>>
>> =================
>>
>>    S57. If C skosxl:hiddenLabel L and L skosxl:literalForm V, then C
>> skos:hiddenLabel V.
>>
>>    <rdf:Description>
>>        <rdfs:subPropertyOf
>> rdf:resource="http://www.w3.org/2004/02/skos/core#hiddenLabel"/>
>>        <owl:propertyChain rdf:parseType="Collection">
>>            <rdf:Description
>> rdf:about="http://www.w3.org/2008/05/skos-xl#hiddenLabel"/>
>>            <rdf:Description
>> rdf:about="http://www.w3.org/2008/05/skos-xl#literalForm"/>
>>        </owl:propertyChain>
>>    </rdf:Description>
>>
>> ==============
>>
>>
>> [1] http://www.w3.org/2006/07/SWD/SKOS/reference/20081001/skos.rdf
>> [2] http://www.w3.org/2006/07/SWD/SKOS/reference/20081001/skos-xl.rdf
>> [3] http://www.w3.org/TR/2008/WD-rdf-text-20081202/
>> [4] http://www.w3.org/TR/2008/WD-owl2-quick-reference-20081202/
>> [5] http://www.w3.org/TR/2008/WD-owl2-mapping-to-rdf-20081202/
>>
>>
> 
Received on Sunday, 22 March 2009 16:18:22 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Sunday, 22 March 2009 16:18:23 GMT