- From: David Powell <djpowell@djpowell.net>
- Date: Wed, 16 Mar 2005 23:45:09 +0000
- To: "Joshua Allen" <joshuaa@microsoft.com>
- CC: "Geoff Chappell" <geoff@sover.net>, "Danny Ayers" <danny.ayers@gmail.com>, "Seth Russell" <russell.seth@gmail.com>, <semantic-web@w3.org>
Hello Joshua, Wednesday, March 16, 2005, 11:02:20 PM, joshuaa@microsoft.com wrote in mid:0E36FD96D96FCA4AA8E8F2D199320E5204883C09@RED-MSG-43.redmond.corp.microsoft.com : >> > unspecified? And of course, all of it means I *need* RDFS (and > OWL), >> > *everywhere*. Why take such a huge dependency if you don't need to? >> >> So if you have 8 boolean properties, you need a minimum of 8 classes > (or > If I use Boolean URIs, I don't need *any* classes. > I'm struggling to see the value in modeling this in any way that > requires more than zero classes (or OWL) to be defined. Using: <http://foobar/page.html> <rdf:type> <urn:myterms:CachedObject> isn't any more difficult that using: <http://foobar/page.html> <urn:myterms:isCached> <http://xmlns.com/wordnet/1.6/true> The class <urn:myterms:CachedObject> is just a resource in the same way as a Boolean URI such as <http://xmlns.com/wordnet/1.6/true> is a resource. It would be good practice to declare the class in some RDFS, but it wouldn't be wrong if you don't. In fact, if you were using an RDFS reasoner, then I guess that the presence of an rdf:type property in your graph would implicitly declare the class due to the range of the rdf:type property being rdfs:Class. Using classes to represent boolean properties seems like a better RDF idiom than using boolean properties. It also makes it easier to document the properties using RDFS if you want to. -- Dave
Received on Wednesday, 16 March 2005 23:45:42 UTC