- From: Xiaoshu Wang <wangxiao@musc.edu>
- Date: Thu, 19 Apr 2007 11:57:55 -0400
- To: Bijan Parsia <bparsia@cs.man.ac.uk>
- CC: Alan Ruttenberg <alanruttenberg@gmail.com>, Jeremy Carroll <jjc@hpl.hp.com>, public-owl-dev@w3.org, Tim Berners-Lee <timbl@w3.org>, Alistair Miles <a.j.miles@rl.ac.uk>
Bijan, > That also is specious, IMHO. First it conflates two things, 1) whether > there is an autoinclude or a explicit include mechanism and 2) whether > owl:imports can be legally used in this case. > > In point of fact, it can with no discernable difficulty. And even if > it couldn't, that doesn't preclude an explicit model. Indeed, it > strongly supports it (i.e., merging graphs is outside standard RDF). I didn't say it can't. I was just saying, if there is auto-include, what is the point of explicit include? What kind of difference does it make? > So you are free to follow your nose in your tools, but if I am using > your tools I shall complain unless there is an opt out by default. > >> (If it is, then what isn't?) > > I'm fine with every rdf graph being an owl:Ontology. In point of fact, > they are (at least owl full). So that's fine. I totally agree that every rdf graph is an ontology. But then, shouldn't be owl:imports be relaxed to not constraining its domain to owl:Ontology. And, then, what is the point of owl:Ontology? Shouldn't it be removed? > Second, your example contains a subclass statement which is generally > an ontology statement. So this is a *good* example of an ontology (if > small). Of course, I consider http://example.com/o1 an ontology. That is not the question, the question is if the statement _:x a http://example.com/o1#A is an Ontology? (I cannot do the owl:imports here but not at http://example.com/o1) >> And since there is no RDF/RDFS vocabularies that are given an >> "import/include" role, the only logical conclusion is that RDF must >> be using a "nose" model. > > Far from being the *only* logical conclusion, it's a non sequitur > (that is, not even logical). > >> This also makes sense since one of the essential features of the web >> is "self-descriptive" and a "nose" model fits in satisfactorily. > > Non sequitur number 2! Sequitur or not. :-) I am still not sure what is your position on the answer to my example? Fact: _:x a http://example.com/o1#A Question: _:x a http://example.com/o2#B? What should be the desired answer in semantic web? Similarly, if I say to you that "I am a person"? What will be your answer to the question if I am a mammal? Now, imaging two machine agents are communicating? What does it means if agent A said to agent B that _:x a http://example.com/o1#A? Not much beyond this sentence or potentially a lot more? > Now you're being prescriptive. A conforment reasoner *must not*, in > it's conforming behavior, do what you claim it has to. That doesn't > preclude a larger *system* from doing this, to my mind, typically > useless, behavior. But if there is no standard behavior prescribed, how do we communicate with clarity? I thought the whole point of semantic web is to be explicit and unambiguous, right? > If I use an URI in an RDF or OWL document I publish, I intend nothing > more to get into that document than the what's in the imports closure. > If you add extra stuff that's your look out. I don't have problem with explicit import model. The problem that I have is the ambiguity in the current RDF and OWL spec. First, standard "import" behavior needs to be explicitly described. Second, depending on how the above question is answered, owl:imports, (perhaps, owl:Ontology and OntologyProperty as well) needs to be revised accordingly. Cheers, Xiaoshu
Received on Thursday, 19 April 2007 16:01:34 UTC