RE: On the integration of Topic Maps and RDF

Hi Eric,

thanks very much for your comments !
I' glad you liked our paper.

> It seems to me there could be a couple different interpretations in
> modeling topic maps in this way.  Have you received any feedback from the
> Topic Map community on your particular representation?

I just had a series of long discussions with the Topic Map community at a
conference last week, where I also presented our approach. I had a lot of
positive feedback. The Topic Map commmunity is well aware of the potential
of RDF for the Semantic Web and is very interested in using RDF as the basis
for Topic Maps - provided a more mature and formal representation.
Also, the core ISO standard is work in progress. In the near future, it will
include the grapph-based data model which we use as the basis for our
mapping. Great news for interoperability with RDF.

You mean the interpretation of our way to model Topic Maps is ambiguous ?
There is one missing piece and that is a set of rules which would allow the
user queries for "higher-level" Topic Map constructs instead of RDF path
expressions. These rules could be expressed in an RDF rule language which
allows to specify data locations including their respective data model.

There are several other ways to model Topic Maps with RDF. These have been
nicely summarized by Graham Moore in his XML 2000 paper. Nikita Ogievetsky
also presented a way to go directly from the XTM syntax to an RDF syntax
with an XSLT stylesheet.

Issues of discussion included also in what way the proposed graph model
(i.e. the types of arcs and nodes) of Topic Maps is minimal or could be
replaced by a smaller (in the set of primitives) RDF model. A big issue for
the Topic Map community seems to be the lack of distinction between what
they call "subject indicating resource" and "subject constituting resource"
in RDF. I am not sure yet, whether this is an issue for RDF or not.

Cheers,

Martin

Received on Tuesday, 21 August 2001 13:44:33 UTC