- From: Danny Ayers <danny.ayers@gmail.com>
- Date: Wed, 25 Nov 2009 01:52:24 +0100
- To: martin.hepp@ebusiness-unibw.org
- Cc: Paul Houle <ontology2@gmail.com>, nathan@webr3.org, pedantic-web@googlegroups.com, Linked Data community <public-lod@w3.org>, Alison Broady <alib@wyehead.co.uk>
Hi Nathan, A good question, the way it gets answered as far as I can see depends on what you're after. Glad to see you're thinking linked data. But people really do try to overthink it when it comes to ontologies, in my opinion: ideally the best ontologies/vocabs will win - - rubbish. The ontologies/vocabularies on machines will always be poor reflections of the things they try to describe. There is a huge amount of software around these days (and has been for many years) that tries to describe things. The advantage that the Web languages have is that it can work in a big distributed environment. Put a marker down (a URI) for a concept or a dog and it's reusable. When it comes to multiple ontologies - yes, it's a reality. In practice maybe it means lots of different clauses in the query - but that depends on how far you want to ask. What are you interested in? Certainly good practice says as a publisher of information you should use existing terms wherever appropriate rather than invent (c'mon, should I call the creature next to me a wurble or a dog?). But everyone can make up their own terms, and there's nothing wrong with that. To answer your subject line, the only way we can avoid ontology wars is by making the field flat, globally. I think we have that now, at least in principle. Cheers, Danny. -- http://danny.ayers.name
Received on Wednesday, 25 November 2009 00:52:57 UTC