[rdfweb-dev] Extending a schema

Sebastian Dietzold sebastian.dietzold at imise.uni-leipzig.de
Tue Mar 9 02:22:48 UTC 2004


quote Ian Davis (8.3.2004):

> > Now, that validates fine, but when I look at the graph I see an extra
> > duplicated "type" edge bewteen mbox and "property"  Is that something
> > I should be worried about?  I assume it would not be proper to take
> > the existing FOAF.rdfs and add my stuff inline with it.
> I wonder what software you're using to view the graph. I suspect that
> the extra Property arc is being generated from the rdfs:domain of your
> longname property.

Hello Ian and Alex, the reason of the edge between foaf:mbox and
rdf:Property is the description of foaf:mbox, not the rdfs:domain of
ui:longname.

The XML code:

<rdf:Property rdf:about="http://xmlns.com/foaf/0.1/mbox">
  <ui:longname>Primary Email Address</ui:longname>
</rdf:Property>

provides two tripel (i assume typical namespaces):

foaf:mbox rdf:type rdf:Property
foaf:mbox ui:longname "Primary ..."

i think alex should write instead:

<rdf:Description rdf:about="http://xmlns.com/foaf/0.1/mbox">
  <ui:longname>Primary Email Address</ui:longname>
</rdf:Description>

This announce only one statement and another advantage is, that in a
change of foaf:mbox (maybe from a property to class, we got a message
because the domain stipulation of ui:longname is not fulfilled anymore.

> It's not incorrect to have duplicate triples but sometimes confusing to
> display them.

Its not incorrect but imho we should take care of making to much
statements about foreign resources. its a critical thing to use other
properties than annotation-prop, cause you dont really know what the
resource is and how it change.

hope, i could help.

   Sebastian Dietzold

-- 
Sebastian Dietzold
Content Management / Directory Services
Institute for Medical Informatics, Statistics and Epidemiology (IMISE)
University of Leipzig; Liebigstr. 27; 04103 Leipzig
Phone: +49 341 97 161 14; Fax: +49 341 97 161 30



More information about the foaf-dev mailing list