- From: Peter Mika <pmika@cs.vu.nl>
- Date: Fri, 4 Jun 2004 12:15:25 +0200
- To: "'Natasha Noy'" <noy@SMI.Stanford.EDU>
- Cc: <public-swbp-wg@w3.org>
Dear Natasha, (cc: SWBP-WG) > Or OWL Full for that matter. However, this is an interesting point, and > perhaps adding this possibility as one of the bullet points for > approaches 2 and 3 would make sense. I'll do that for the next version > and see what other say. It can be expressed in SWRL (Patel-Schneider and Horrocks, WWW2004). But if you don't want to commit to any particular representation (since that game is not over yet...), then you might choose to stay with a plain old FOL formula :) BTW, it seems like Approach 3 could also use some rule to keep the two hierarchies consistent. (However, this rule cannot be defined without committing to a particular parentSubject property and then better left to the user...) > You have a point there, and others have raised similar issues [1], [2]. > I am not sure (1) is a feasible option though -- see my reply to Brian > today. On (2) indeed translations between patterns would be helpful. > Short of providing the translation tools themselves what type of > information should we provide in the note to enable tool developers to > write conversions? Let me just brainstorm aloud and correct me later if my logic goes astray somewhere. We have five ontologies, which potentially means 4*5/2=10 mappings to do... Alternatively, we could map all of them to a selected approach (to the _least_ expressive one?) or to a foundational ontology that captures concepts on a higher level than the individual ontologies themselves. Aldo and I have used the foundational approach in the past, but it carries a lot of commitment, so now let's just look if it's possible to find internal mappings. For the sake of argument, I'll try to map Approach 2-5 to Approach 1. (Simply by looking at the number of boxes, Approach 1 seems to be the least expressive, which holds the promise that we can project the other ontologies onto this one.) Note: I'll have to index the dc:subject property with the number of the approach, since it is simply not possible that all approaches are talking about the same dc:subject property... For one, they do range on very different things. Note also that I don't introduce any type restrictions; I guess this document would not want to mandate the use of, for example, a certain Subject concept as a top level class of all subjects. Mappings: Approach 2 -> 1: (x, dc:subject_2, y) AND (y, rdf:type, z) -> (x, dc:subject_1, z) Approach 3 -> 1: (x, dc:subject_3, y) AND (y, rdfs:seeAlso, z) -> (x, dc:subject_1, z) Approach 4 -> 1: Seems to be the same as 2 -> 1, but again a different use of dc:subject (one that ranges on Lions and not Subjects. (Without foundations, we will never resolve this distinction, however.) Approach 5 -> 1: Same as Approach 1, except that dc:subject_5 is an annotation property, which puts limitations on it when looked at it as OWL DL. In summary: such mappings are possible, at least for translating the OWL DL approaches to the simplest OWL Full version. The translation is not expressible in OWL Full, however. Also, there is a need to distinguish the five separate uses of dc:subject, which might actually be the solution for the question of how a machine could tell the different approaches apart. The other direction of translation (not presented here) seems also feasible, even if with some existential quantifications. Best, Peter > Natasha > > [1] http://lists.w3.org/Archives/Public/public-swbp-wg/2004May/0105.html > > [2]http://lists.w3.org/Archives/Public/public-swbp-wg/2004May/0124.html
Received on Friday, 4 June 2004 06:16:00 UTC