- From: Jeremy Carroll <jjc@hpl.hp.com>
- Date: Sun, 26 Jan 2003 17:37:10 +0100
- To: www-webont-wg@w3.org
Can someone point me at an example of the use of owls:Documentation, Most of Jérôme's message seemed related to the topic of annotations of one sort or another - I thought I might throw into the pot an observation or two: rdfs:label is important, partly because it is more-or-less specifically mentioned in our requirement "R16. User-displayable labels" [1] An ontology like: http://mnemosyne.umd.edu/~aelkiss/weather-ont.daml shows a use of how we meet this requirement with rdfs:label. Some of it is: <WeatherIntensity rdf:ID="Medium"> <rdfs:label>Medium Intensity</rdfs:label> </WeatherIntensity> ... <daml:Class rdf:ID="WeatherIntensity"> <rdfs:label>Weather Intensity Descriptor</rdfs:label> <rdfs:subClassOf rdf:resource="#WeatherQualifier"/> </daml:Class> (the objects appear to be sorted alphabetically by fragID). I note this usage is currently not supported in OWL DL or OWL Lite; because the 'annotation' is being applied to both a class in the ontology and an individual. Given this labelling of resources in an ontology is a requirement, I wonder whether this simple unproblematic usage could instead be blessed. Minimally one could allow rdfs:label and rdfs:comment on any resource in the ontology; although it seems more natural to allow user defined annotations on classes, properties and ontologies as well. The main reason rdfs:label is unproblematic in this ontology is that it does not interact with the class hierarchy - there are no domain or range constraints on it, nor any restrictions using it, nor is it a subproperty of a property with such features. (There is the range constraint on it that its range is rdfs:Literal, this is essentially equivalent to declaring it to be of owl:DatatypeProperty). Calling such unproblematic properties as simple properties, I do not see why we do not just use simple datatype properties and simple object properties uniformly as annotations and also permitted to be used with individuals in the domain of discourse. In either case the meaning is essentially that defined by RDFS (not hugely interesting). Jeremy [1] http://lists.w3.org/Archives/Public/www-webont-wg/2003Jan/att-0347/01-webont-req#req-user-labels
Received on Sunday, 26 January 2003 11:36:20 UTC