- From: Jan Wielemaker <jan@swi.psy.uva.nl>
- Date: Fri, 6 Dec 2002 14:35:57 +0100
- To: www-rdf-interest@w3.org
Hi,
[ hope this doesn't get twice to the list; I found the dialog to publish
a bit misleading: I wanted to grant permission for everything and then
it tells me th message will *not* be distributed
]
I'm working on a facelift of the SWI-Prolog RDF infrastructure. The
parser now (almost, there are some dubious tests) conforms to the RDF
revised syntax specs.
Based on some practical experience how to use RDF I'm now redesigning
the storage and query interfaces. In our projects we've seen the use
of the sub-properties is a very neat way to relate information from
multiple sources created with different viewpoints, so handling this
information will now be a basic property of the query engine.
How far however should it go? Can someone make a sub-property if
rdf:type and claim that a new subject with this property belongs to
the specified type?
Given the fact that rdf:type is part of base RDF, where sub-properties
are part of RDFS suggests this might not be a good idea. It might also
seriously harm performance as adding/deleting such a subproperty relation
can have huge consequences.
Opinions?
--- Jan
Received on Friday, 6 December 2002 08:42:37 UTC