- From: Jeff Heflin <heflin@cse.lehigh.edu>
- Date: Fri, 20 Sep 2002 10:51:13 -0400
- To: Jim Hendler <hendler@cs.umd.edu>
- CC: www-webont-wg@w3.org
Jim Hendler wrote: > > At 5:42 PM -0400 9/19/02, Jeff Heflin wrote: > >Chris, > > > >Just to refresh everyone's memory, this discussion began because of the > >proposals for handling imports that are based in XML/RDF syntactic > >inclusion. My whole point is that in every approach I've seen so far, > >this form of inclusion loses track of which statements come from which > >ontologies. In these cases, I no longer have the URLs of the documents; > >they were lost when one document was inserted into another. If all OWL > >parsers must perform this kind of inclusion, then people like me lose > >the ability to do the things I mentioned. > > > >Note, I was not arguing that we need to modify RDF or OWL in some > >radical way so that the model inherently includes source information > >(I'll save that battle for OWL 2.0 :-). I simply want to make sure we > >have solutions that don't prevent people who need source information > >from doing what they want. > > > >Jeff > > Jeff - wouldn't rdfs:IsDefinedBy [1] work? - JH The problem with IsDefinedBy is that it doesn't even make sense in RDF. On the Semantic Web, there is no single place to "define" a resource. The "definition" could be any number of RDF statements distributed among many web pages. If you care about trust and things like that, then you need to be able to to say which of the "definitional" statements come from which ontologies. Currently, the only way to do that in RDF is to use reification, as Jonathan has pointed out. However, we also have the option to do it outside of RDF as Dan Connolly does (by simply keeping track of URLs in his applications). Believe it or not, I'm actually okay with this approach for now (we'll have to fix the mess when we get to the trust level, but let's let another WG worry about that). I only brought the point up because I believe that approaches which rely on textual inclusion of XML syntax will prevent me from figuring out the source of a statement because once you do the inclusion, you no longer have the URLs of the individual files. For example, if we use the XInclude approach, then an XML parser will output a single DOM that merges the DOMs of the importing and imported ontology. In the process, we have lost information (the sources of the various RDF statements) that would have been valuable to my application. In short, as long as we don't do an RDF/XML syntactic inclusion method (that is option d to Question 1 from [1]), I believe my need for "knowing where things come from" can be satisfied. If we can drop that option from the table, then I see no point in arguing this issue further. Jeff [1] http://lists.w3.org/Archives/Public/www-webont-wg/2002Sep/0272.html > [1] http://www.w3.org/TR/rdf-schema/#ch_isdefinedby > -- > Professor James Hendler hendler@cs.umd.edu > Director, Semantic Web and Agent Technologies 301-405-2696 > Maryland Information and Network Dynamics Lab. 301-405-6707 (Fax) > Univ of Maryland, College Park, MD 20742 240-731-3822 (Cell) > http://www.cs.umd.edu/users/hendler
Received on Friday, 20 September 2002 10:51:16 UTC