Hi all > Note that the dbpedia ontology isn't really an ontology of the > world, it's really an ontology about wikipedia entries > Indeed. And it should stay that way, and of course it is bound to be globally inconsistent forever, as Wikipedia is and will continue to be. There is no way it could be otherwise, since collective knowledge is inconsistent passed a certain scope and size. And this is IMHO rather a good thing. The perspective of a globally consistent view of the world is really a totalitarian nightmare! Even each of our individual knowledge(s) is inconsistent, I guess. At least so is mine. :-) So what is DBpedia ontology good for? Obviously for (partial) use in data linking and integration, but certainly not for building a global consistent view of the world à la Cyc. But a potential benefit of reasoning on this ontology, and maybe the most interesting one in the long run, is to explicit inconsistencies and feed them back to Wikipedia curators (as has been said, with all due diplomacy to Wikipedia work). Editors down there are often very well aware of inconsistencies, they are often pointed in discussion pages, for instance, decisions about it might be pending due to edition wars etc. But pinpointing them by expliciting the semantics does help. Some inconsistencies can be obvious errors or categorisation, like putting a place in a "person" category, but I guess many (most) of them actually point to things not yet clearly defined, or about which conflicting ontologies coexist. In such cases, they are topics about which knowledge is in the making, very often the most interesting ones! Maybe it could be suggested to Wikipedians to create a category "Logically Inconsistent", to mark those articles and categories. This shows at Web scale what I experience daily in data migration to semantic environments. Even if, as often happens, the resulting system do not meet the initial expectations, a side effect is in any case to have people look closely at their data and vocabulary for the necessary semantic sanity check, and this is very useful per se. If you ask me, even more useful than the system implemented, which is bound to have a very short life cycle anyway, as we all know. As Bruno Bachimont uses to say, an ontology is mainly a tool to explicit inconsistencies of our knowledge, pointing to new questions for research. After that, you can throw it away. :-) When the ontology is consistent, you have nothing to discover any more. Trying by all means to hide or wipe out inconsistencies is based on some paranoid logician's view of the world, similar to those who did not want to look into Galileo's telescope because sunspots, Moon's mountains and Jupiter's satellites were not mentioned in Aristote and the Bible ...
Cheers -- *Bernard Vatant *Senior Consultant Vocabulary & Data Engineering Tel: +33 (0) 971 488 459 Mail: [email protected] <mailto:[email protected]> ---------------------------------------------------- *Mondeca** *3, cité Nollez 75018 Paris France Web: www.mondeca.com <http://www.mondeca.com> Blog: Leçons de Choses <http://mondeca.wordpress.com/> ----------------------------------------------------** ------------------------------------------------------------------------------ Let Crystal Reports handle the reporting - Free Crystal Reports 2008 30-Day trial. Simplify your report design, integration and deployment - and focus on what you do best, core application coding. Discover what's new with Crystal Reports now. http://p.sf.net/sfu/bobj-july _______________________________________________ Dbpedia-discussion mailing list [email protected] https://lists.sourceforge.net/lists/listinfo/dbpedia-discussion
