Michael, Thanks for the reference and comment. That does answer part of the question I had. This still leaves me wondering whether the DBpedia ontology is being intentionally kept small (rather than representing all of the classes described in Wikipedia) for performance reasons, or because the effort of trying to make the ontology accurate while expanding it substantially is too large for the existing team. If the latter, perhaps I can help in that task. If there are performance issues, I would like to get more detail to see just how serious an expansion of the ontology would be for the existing usages.
The other issue raised by the discussion in the FAQ at the GoodRelations page you cited is - just what should the semantics of a Wikipedia article be? At present, the "type" relation is used, and that make it appear odd for traditional ontology structure, when the Wikipedia article describes a class of things rather than an individual. It seems (to me) that it would be more appropriate to use a relation of the type "isaDiscussionOf" (where the main topic of the article is exactly the same as the intended class of the Wikipedia), or "isaDiscussionOfaSubtypeOf" where there is no corresponding DBpedia ontology entry for that class, but the class described can be related to a broader class in the ontology. The "type" relation might still be used, but to be consistent with the notion of Wikipedia articles being discussions rather than the individuals described it might be better to use a relation such as "isDiscussionOfanInstanceOf" to point to the class of the entity discussed. Then, the articles would all have the semantics of articles rather than of the entities described by the articles. I do recall someone (not sure who) mentioning that this issue of the proper semantics for articles has been an issue, but I don't recall in what context. I suppose it may well have been a hot topic for the DBpedia community at some time. I apologize that I am quite new to this community, so do not know the history. But unless there are performance issues that are truly insurmountable, it seems to me not to be difficult to get the semantics right so that the ontology can grow to accurately reflect properties and relations of the entities discussed by Wikipedia. And I will be willing to help in that task. Or . . . is this a task that has consciously been left to the GoodRelations group? Pat Patrick Cassidy MICRA Inc. cass...@micra.com 908-561-3416 -----Original Message----- From: Michael Brunnbauer [mailto:bru...@netestate.de] Sent: Sunday, January 01, 2012 10:35 AM To: Patrick Cassidy Cc: dbpedia-discussion@lists.sourceforge.net Subject: Re: [Dbpedia-discussion] DBpedia ontology Hello Pat, DBpedia is not an ontology - it is a dataset of instance data that partly uses an ontology created by the DBpedia team. The Instance data (ABox) has the namespace http://dbpedia.org/resource/ - e.g. http://dbpedia.org/resource/Albatross The Ontology (TBox) has the namespace http://dbpedia.org/ontology/ - e.g. http://dbpedia.org/ontology/Bird I agree with you that it would be nice if many DBpedia entries would be classes with proper subclassing instead of individuals. But as DBpedia data is generated automatically from Wikipedia, I guess it would be quite difficult to keep this consistent. You are not the first one to notice these deficiencies in the use of DBpedia, see "Frequently Asked Questions" on http://www.productontology.org/ Regards, Michael Brunnbauer ------------------------------------------------------------------------------ Ridiculously easy VDI. With Citrix VDI-in-a-Box, you don't need a complex infrastructure or vast IT resources to deliver seamless, secure access to virtual desktops. With this all-in-one solution, easily deploy virtual desktops for less than the cost of PCs and save 60% on VDI infrastructure costs. Try it free! http://p.sf.net/sfu/Citrix-VDIinabox _______________________________________________ Dbpedia-discussion mailing list Dbpedia-discussion@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/dbpedia-discussion
