- From: Michael Schneider <schneid@fzi.de>
- Date: Tue, 21 Aug 2007 20:42:21 +0200
- To: <reto@gmuer.ch>
- Cc: <semantic-web@w3.org>
Hi Reto! Reto Bachmann-Gmür <reto@gmuer.ch> wrote: > I'm writing an ontology with properties pointing to literals > representing the following: > - media types (eg: "application/xhtml+xml") > - media ranges (eg: "application/*") > - fully qualified java-class names (eg: "org.example.tools.MyTool") > > While I could use xsd:string I think it would be better to use custom > datatypes to have literals like > "org.example.tools.MyTool"^^http://example.org/datatypes#javaClass. > > My question: > - Anyone knows a place where such datatypes are already defined? > - Datatypes are typically defined in an XML-schema, is there an ontology > to do so in RDF? > - Is it a good idea to define the datatypes in the same namespace as the > ontology using them? I'm afraid that I cannot answer your questions here. I have never worked with custom datatypes in RDF, or even created any. And I would surly be reluctant to do so, before not having considered other possibilities. Thus my question: Are you really restricted to datavalues for your application? I could imagine an alternative approach, which might or might not serve your needs. Instead of encoding the mime types and java class names into data literals, you could define URIs for them in the following way: * urn:mime:application/xhtml+xml * urn:java:org.example.tools.myTool Note here, that these URIs are really unique in each case, due to the unique naming schemes for encoding mime types and Java class names. Then, instead of specifying custom datatypes, you can define OWL classes, and assign them to your URIs: ex:MimeType a owl:Class . <urn:mime:application/xhtml+xml> a ex:MimeType . ex:JavaClass a owl:Class . <urn:java:org.example.tools.myTool> a ex:JavaClass . I do not know your concrete application, but when you say that you use "properties pointing to literals", I would believe that you intent to use the different kinds of literals for annotating resources. Perhaps something like multimedia stuff, which is annotated by its mime type and the decoder pluging class to use? You can do this with my idea, too (using object properties instead of datatype properties): <http://videobase.org/film/kubrik/2001> ex:hasMimeType <urn:mime:video/mpg4> ; ex:hasPlugin <urn:java:org.videoplugins.decoder.MP4> . or something similar (I just made this up). Maybe, there are alread URN schemes or ontologies for mime types and java class names, but I don't know. Now, can you live with such an alternative approach? Cheers, Michael -- Dipl.-Inform. Michael Schneider FZI Forschungszentrum Informatik Karlsruhe Abtl. Information Process Engineering (IPE) Tel : +49-721-9654-726 Fax : +49-721-9654-727 Email: Michael.Schneider@fzi.de Web : http://www.fzi.de/ipe/eng/mitarbeiter.php?id=555 FZI Forschungszentrum Informatik an der Universität Karlsruhe Haid-und-Neu-Str. 10-14, D-76131 Karlsruhe Tel.: +49-721-9654-0, Fax: +49-721-9654-959 Stiftung des bürgerlichen Rechts Az: 14-0563.1 Regierungspräsidium Karlsruhe Vorstand: Rüdiger Dillmann, Michael Flor, Jivka Ovtcharova, Rudi Studer Vorsitzender des Kuratoriums: Ministerialdirigent Günther Leßnerkraus
Received on Wednesday, 22 August 2007 03:58:34 UTC