RE: [uddi-spec] OWL UDDI

["John Colgrave" <colgrave@hursley.ibm.com>]

There is no such thing as "OWL UDDI" now.  The decision we took in NZ was to
require a UDDI implementation to accept any OWL ontology but we would
describe what subset of the ontology contributed to the UDDI taxonomy.  This
subset could be regarded as a kind of "OWL UDDI" but it should not be
thought of as restricting what features of OWL can be used in the original
ontology.  As part of the rewrite I am describing these contributions in
terms of triples rather than OWL/RDF/XML syntax.

I don't understand your comment about section 2.1.4 and I don't think your
proposed changes have any effect.

Section 2.1.4 does not attempt to redefine owl:Class, it simply defines an
owl:FunctionalProperty that has a domain of owl:Class.

Your proposal would require OWL ontologies to be written specifically for
UDDI, as they would have to use UDDI-Class instead of the standard
owl:Class, and would still require OWL full as you use UDDI-Class as the
domain of the property, and it is this that makes it OWL Full.

John Colgrave
IBM


> -----Original Message-----
> From: Massimo Paolucci [mailto:paolucci@cs.cmu.edu]
> Sent: 05 August 2004 22:41
> To: uddi-spec@lists.oasis-open.org
> Subject: [uddi-spec] OWL UDDI
> 
> This message is in reference to the status of OWL UDDI with respect to
> the hierarchy of OWL languages (Lite/DL/Full).  In the UDDI spec TC V4
> Proposal: OWL as UDDI Taxonomy Language it is claimed that OWL UDDI is
> OWL Full.  Actually I think that OWL UDDI can very easily be re-casted
> in OWL Lite,  furthermore I suspect that the definition presented in
> section  2.1.4 is actually wrong because it tries to redefine the basic
> OWL class owl:Class.
> 
> What I propose instead is to redefine OWL UDDI on the bases of the
> following class.
> 
>   <owl:Class rdf:ID="UDDI-Class">
>     <rdfs:label>UDDI-Class</rdfs:label>
>     <comment>
>       Class that specifies the types of classes that can be used in UDDI.
>       This class is characterized of one (optional) boolean property
> called
>       "selectable", characterized by a cardinality 1 restriction.
>     <comment>
>     <rdfs:subClassOf>
>     <owl:Restriction>
>       <owl:onProperty rdf:resource="selectable"/>
>       <owl:maxCardinality
> rdf:datatype="&xsd;nonNegativeInteger">1</owl:maxCardinality>
>     </owl:Restriction>
>     </rdfs:subClassOf>
>   </owl:Class>
> 
>   <owl:DatatypeProperty rdf:ID="selectable">
>     <rdfs:domain rdf:resource="http://www.w3.org/2002/07/owl#UDDI-Class"/>
>     <rdfs:range rdf:resource="http://www.w3.org/2001/XMLSchema#boolean"/>
>   </owl:DatatypeProperty>
> 
> If we adopt this view, OWL UDDI is just OWL Lite  (cardinality
> restriction of at most 1).
> 
> the assumption that any class loaded in UDDI either has the selectable
> property set to false or it is unselectable would still hold, in
> addition given a taxonomy, to make any class unselectable it is enough
> to add an about statements like in the following example:
> 
>   <owl:Class rdf:about="http://www.unspsc.org/UNv61101#_90121602";>
>     <rdfs:subClassOf rdf:resource="http://uddi.org/owl#UDDI-Class"/>
>     <uddi:selectable
> rdf:datatype="http://www.w3.org/2001/XMLSchema#boolean";>false</uddi:select
> able>
>   </owl:Class>
> 
> If this solution is adopted, the TN would need some minor editing to
> section 2.1.4 and to the example taxonomies.  I will send the corrected
> TN tomorrow
> 
> --- Massimo
> 
> 
> To unsubscribe from this mailing list (and be removed from the roster of
> the OASIS TC), go to http://www.oasis-open.org/apps/org/workgroup/uddi-
> spec/members/leave_workgroup.php.