Forwarded message 1
Dear Guha,
At WWW'02, after your presentation, I told you I had converted WordNet 1.7
into an ontology usable for knowledge representation, and that this work
seems complementary to TAP since the TAP vocabulary seems too small to permit
knowledge sharing (I agree that the number of "classes and property types
(like temperature and birthplace) is relatively small, at least compared to
the number of nodes" but I do think it still needs to be large enough to
cover all categories referred by common English nouns to permit knowledge
representation (even "to provide descriptions of most of the items on sale")
and knowledge sharing.
This work is now documented and accessible at
http://meganesia.int.gu.edu.au/~phmartin/WebKB/doc/wn/wnIntegration.html
It involved
- the generation of unique and people-friendly identifiers for WordNet categories
- the (re-)organization of WordNet top layers via the insertion of top-level
categories into a top-level ontology
- 296 semantic corrections, 155 category additions, 335 lexical corrections
- the distinction of WordNet generalization links into subtype links and
instance links (6211 individuals were isolated).
I have recently read in http://rdfweb.org/2002/tap/ that you are using
WordNet too, but no details are provided. Did you generate category
identifiers too?
I have also connected the media objects of TAP to WordNet categories:
http://meganesia.int.gu.edu.au/~phmartin/WebKB/kb/tapMediaObject.html
(this file is both a documentation and a file executable by WebKB).
Initially, I intended to integrate all TAP into the KB of WebKB-2.
However, I now do not see the need for storing so many "data" in a KB.
It does not ease or guide knowledge representation, and to really permit
knowledge sharing along your approach, much, much more data is needed
(all books/films/characters/documents ... in the world?). Since,
I already have a vocabulary for commonsense categories, I fail to see the
need. Hence, I will restrict the integration of TAP into WebKB-2 to the
classes (I do not currently think I will integrate the properties because
WebKB-2 offer more precise and normalized ways to represent relations
using both the relation type ontology and the concept type ontology).
If you are interested, I'll send you the result.
Finally, I'd like to report small problems in the RDF file for media objects:
http://meganesia.int.gu.edu.au/~phmartin/WebKB/kb/tapRDFproblems.html
I hope you will view this work as a contribution to TAP.
As I am going to conferences and vacation next week (for 2 months), I intend
to advertize this work on the CG and SUO mailing lists this week.
If you have some comments or corrections on my work, please tell me.
Cheers,
Philippe
_____________________________________________________________________________
Dr. Philippe Martin
Senior Research Fellow at the Distributed Systems Technology Center
(DSTC Pty Ltd; DSTC is W3C's Australian Office)
Address: Griffith Uni, School of I.T., PMB 50 GCMC, QLD 9726 Australia
Email: philippe.martin@gu.edu.au; Fax: +61 7 5594 8066
_____________________________________________________________________________