- From: Enrico Franconi <franconi@inf.unibz.it>
- Date: Sat, 20 Dec 2003 13:35:13 +0100
- To: public-sws-ig@w3.org, www-rdf-logic@w3.org
> [Jim Hendler] > I do think there are some nice doctoral theses to be done (and I'm > hiring students to do them) in automating more of the mapping (human > provides partial map, system extends it) within the sorts of > mappings > that OWL allows, but doing this by humans, using ontology tools, is > presently well within the state of the practice and easy to do .... > > [Drew McDermott] > This I agree with. The point is that once humans have figured out how > two ontologies overlap, it isn't that hard to formalize the > translation. Automated tools can certainly help, and we've been > developing some of them. The DB world is facing this problem since many many years. Look at: E. Rahm and Phil A. Bernstein, "A survey of approaches to automatic schema matching", VLDB Journal 10, 4 (Dec. 2001), at http://www.research.microsoft.com/~philbe/VLDBJ-Dec2001.pdf The two most cited projects about that problem are the DIKE project (L. Palopoli, D. Saccà, G. Terracina, D. Ursino) (http://portal.acm.org/citation.cfm?id=945189&dl=ACM&coll=GUIDE as a starting point) and the Phil Bernstein project at Microsoft.
Received on Saturday, 20 December 2003 07:38:37 UTC