Re: Proposed response for timbl-02 (reification semantics)

* Patrick.Stickler@nokia.com <Patrick.Stickler@nokia.com> [2003-04-10 14:06+0300]
> 
> 
> > -----Original Message-----
> > From: ext Jan Grant [mailto:Jan.Grant@bristol.ac.uk]
> > Sent: 10 April, 2003 13:51
> > To: Stickler Patrick (NMP/Tampere)
> > Cc: danbri; bwm; gk; w3c-rdfcore-wg
> > Subject: RE: Proposed response for timbl-02 (reification semantics)
> > 
> > 
> > On Thu, 10 Apr 2003, Patrick.Stickler wrote:
> > 
> > > > -----Original Message-----
> > > > From: ext Dan Brickley [mailto:danbri@w3.org]
> > 
> > > > here's a sketch towards testcase, sorry haven't polished 
> > this up but
> > > > the intent should hopefully be clear. It takes your sample and
> > > > shows the kinds of interferences that OWL-happy systems 
> > might make in
> > > > the face of knowing some things have multiple URIs and 
> > OWLs ability to
> > > > express equality, directly (sameFooAs) or indirectly
> > > > (InverseFunctionalProperty).
> > > >
> > > > [[
> > > >   _:s rdf:type rdf:Statement ;
> > > >    _:s rdf:subject <s1> ;
> > > >    _:s rdf:predicate <p1> ;
> > > >    _:s rdf:object <o1> ;
> > > >    _:s #source <http://some.server/some_schema.rdf>
> > > >
> > > > <s1> owl:sameIndividualAs <s2>
> > > > <p1> owl:sameIndividualAs <p2>
> > > > <o1> owl:sameIndividualAs <o2>
> > > >
> > > > ...which with OWL semantics I believe gets us to:
> > > >
> > > >    _:s rdf:type rdf:Statement ;
> > > >    _:s rdf:subject <s2> ;
> > > >    _:s rdf:subject <s1> ;
> > > >    _:s rdf:predicate <p2> ;
> > > >    _:s rdf:predicate <p1> ;
> > > >    _:s rdf:object <o2> ;
> > > >    _:s rdf:object <o1> ;
> > > >    _:s #source <http://some.server/some_schema.rdf>
> > > >
> > > >
> > > > which includes the subset of triples,
> > > >
> > > >    _:s rdf:type rdf:Statement ;
> > > >    _:s rdf:subject <s2> ;
> > > >    _:s rdf:predicate <p2> ;
> > > >    _:s rdf:object <o2> ;
> > > >    _:s #source <http://some.server/some_schema.rdf>
> > > >
> > > > which strongly suggests that the triple
> > > >
> > > >    <s2> <p2> <o2> .
> > > >
> > > > ...can be found in some_schema.rdf, rather than 'can be
> > > > deduced from...'.
> > > > ]]
> > 
> > If you want to distinguish between symbols rather than the things they
> > refer to, wouldn't you also balk at
> > 
> > 	<s1> owl:sameIndividualAs <s2>
> > 
> > since they're clearly not the same? 
> 
> I have no problem with <s1> and <s2> denoting the same thing.

me neither, though I agree it can seem an odd way to express things.
OWL, for better or worse, does give us a way to write RDF/XML instance 
data that (when read wearing OWL glasses) tells us that two URIs denote
the same thing. That's just the world RDF lives in now, and the world
reification will be deployed in. 
> 
> > Should we ask that a 
> > health warning
> > be attached to owl:sameIndividualAs?
> 
> Not at all.
> 
> > I'm really concerned about this test case. It seems to mix formal
> > notions (owl:sameIndividualAs) and woolly ones ("strongly suggests").
> > 
> > Presumably there is an analogous warning to be attached to datatypes?
> > Since (sooner or later) we might find the situation that
> > 
> > 	_:jan eg:age "21.0"^^xsd:decimal .
> > 
> > "strongly implies" that

yes (jang), my 'strongly implies' was a bit vague. At that point I 
only had Patrick's property name ('source') to go on. Since then we have
seen the schema, comments on which below. 

> > 
> > 	_:jan eg:age "21"^^xsd:int .
> 
> After thinking about this some more, I've realized that I don't have
> a problem with the inference you suggested. I.e., I'm OK with
> the following entailment
> 
> {
>    _:s a rdf:Statement .
>    _:s rdf:subject <s1> .
>    _:s rdf:predicate <p1> .
>    _:s rdf:object <o1> .
>    _:s #source #x .
>    <s1> owl:sameIndividual <s2> .
>    <p1> owl:sameIndividual <p2> .
>    <o1> owl:sameIndividual <o2> .
> }
> log:implies
> {
>    _:s a rdf:Statement .
>    _:s rdf:subject <s2> .
>    _:s rdf:predicate <p2> .
>    _:s rdf:object <o2> .
>    _:s #source #x .
> }
> 
> because reification, as now defined, is not quoting. 

Ah, ok, that makes things simpler. My previous mail was written in concern that
you somehow wanted to block this inference from going through.

> If we add verbage of any kind, then it should be to the effect
> that such entailments hold.

OK. Patrick, hope you don't mind us picking on you as a sample user of 
rdf reification, but let's follow this one through a bit further.

Your current defintion for 'source' is as follows:

      <rdfs:Property rdf:about="&rdfx;/source">
         <rdfs:comment>An RDF schema in which the statement
occurs.</rdfs:comment>
         <rdfs:domain rdf:resource="&rdf;#Statement"/>
         <rdfs:range rdf:resource="&rdfx;/Schema"/>
      </rdfs:Property>

This (sorry jang)
'strongly suggests'
'seems to me to be saying'
'leads me as an implementor to believe'

that because of 

>    _:s a rdf:Statement .
>    _:s rdf:subject <s2> .
>    _:s rdf:predicate <p2> .
>    _:s rdf:object <o2> .
>    _:s #source #x .

the RDF statement

<s2> <p2> <o2> .

...occurs in the RDF schema referenced.

Now you don't formally define 'occurs' here.  
For the sake of our example scenario, let us assume that <s2>, <p2> and
<o2> URIrefs are not anywhere to be found in the RDF document that
the 'source' property references.

The work of our health warning is to help RDF vocabulary creators use
language in their defintions for properties such as this which won't
raise inaccurate expectations. In this case, an inaccurate expedtation 
would be that the URIs <s2>, <p2> and <o2> are labels on a triple 
from the graph serialized in the RDF schema document referenced by the 
source property. (or some refinement of that; it's hard to word this stuff)

Does that seem correct? ie. that words like 'occurs' in this context
are super-slippery... 

Dan

Received on Thursday, 10 April 2003 07:34:11 UTC