W3C home > Mailing lists > Public > public-powderwg@w3.org > January 2008

Re: status report - formal layer

From: Phil Archer <parcher@icra.org>
Date: Thu, 17 Jan 2008 16:37:08 +0000
Message-ID: <478F8434.1010009@icra.org>
To: Jeremy Carroll <jjc@hpl.hp.com>
CC: public-powderwg@w3.org


Can I just ask a quick question about what you've written in the OWL wiki.

Your Example of POWDER Full is in two parts - one that is an "OWLed up" 
version of the primary example DR and another which expresses the sub 
class relationship. Is your view of POWDER Full (or of a DR-S) that it 
is both of these or just one of them?

It looks to me as if an RDF/XML instance as shown below, which is the 
sub class relationship plus the attributions triples, does the whole 
thing? hence the point about each DR having to exist as a separate document?



  <rdf:Description rdf:about="">
     <foaf:maker rdf:resource="http://authority.example.org/foaf.rdf#me" />
     <dc:description>Textual information to display to end users

  <wdr:URISet rdf:nodeID="URIs">
       <owl:intersectionOf rdf:parseType="Collection">
           <owl:onProperty rdf:resource="&wdr;includeHosts" />

  <owl:Class rdf:nodeID="descriptors">
       <owl:intersectionOf rdf:parseType="Collection">
           <owl:onProperty rdf:resource="&ex;property1" />
           <owl:hasValue>value 1</owl:hasValue>
           <owl:onProperty rdf:resource="&ex;property2" />
           <owl:hasValue>value 2</owl:hasValue>

    <owl:onProperty rdf:resource="&wdr;hasURI"/>
    <owl:someValuesFrom rdf:nodeID="URIs"/>
    <rdfs:subClassOf rdf:nodeID="descriptors"/>


Jeremy Carroll wrote:
> I didn't make much progress on Friday over and above what I've already 
> pointed to.
> http://www.w3.org/2007/OWL/wiki/POWDER
> I need to do other work, and hope to work on the powder stuff again 
> either Thursday or Friday this week.
> On the validity issue, my current thought is as follows.
> - Each DR should be in its own document (at least notionally: i.e. there 
> is a URL that if you do a get you get the DR - that's not to exclude the 
> case where mutliple DRs are listed together in a single file, in 
> addition to the individual files).
> - The validity refers to the validity of the file (and hence of the 
> individual DR).
> - The subclass relationship corresponding to a DR  necessarily follows, 
> even if the DR is invalid (But invalid DRs may or may not be true).
> - To work out the formal meaning of a set of powder files, the first 
> step is to take the RDF merge of the valid files, and then take the RDF 
> semantics for that merge.
> This follows the named graphs paradigm, in leaving 'difficult' stuff 
> (e.g. time) outside the formal treatment, but providing a pragmatic 
> treatment (ignore invalid files).
> I still need to think through the relationship with packages.
> Unless the discussion indicates that the above approach is misguided, 
> I'll write it up, as part of the proposed formal treatment, and include 
> it in the wiki page.
> Jeremy
Received on Thursday, 17 January 2008 16:37:22 UTC

This archive was generated by hypermail 2.3.1 : Tuesday, 6 January 2015 20:06:03 UTC