- From: Al Gilman <Alfred.S.Gilman@IEEE.org>
- Date: Tue, 13 Mar 2007 20:58:49 -0400
- To: Chris Welty <cawelty@gmail.com>
- Cc: w3c-semweb-cg@w3.org, public-swd-wg@w3.org, lisa@ubaccess.com, wai-liaison@w3.org
At 10:58 PM -0500 10 03 2007, Chris Welty wrote: >Al, > >I just read your message and basically have no idea what you are >trying to do. Of the two theories in your note, theory A is >certainly true: "Owl cardinality constraint can only refer to usages >inside RDF. I don't think we can legally use it to constrain >the usage of native mark up which it abstracts." > >OWL cardinality constraints do, indeed, only apply to RDF data. But >I wasn't able to understand theory B at all, nor your example in its >entirety. But another hint: > >"We can't make the population constraint that there has to be an >@aaa:labeledby attribute in the syntax because it cuts across >element types. But we can say in the role template that something >of this nature has to be there, because it is specific to certain >roles. The question is, is it binding?" > >OWL cardinality constraints are not "closed", they do not work like >e.g. multiplicities in UML. You have more of a clue of what we would like to do than you think. > It is actually quite difficult to violate a minimum cardinality >constraint because the semantics says that if enough values aren't >there then they will come along eventually, and everything else >proceeds as if the values are there. >It's rather like saying "everybody loves somebody" - you don't have >to know precisely who it is that each person loves, just that for >each person the other exists. Thus if you want to use an OWL >cardinality constraint as a sort of syntactic restriction so you can >"reject" a document that doesn't have enough of something, this will >not get you there. Let me parrot back what I think I hear in other words. You can never declare an information pattern scoped to one XHTML document to be deficient based on comparing it with an OWL class, because if some third party elsewhere references the home node in the pattern with an rdf:about in a triple that supplies the missing property, then voila!, the class is fully populated. > >Hope this helps. That helps a lot, if I follow you correctly. This doesn't necessarily end the stream of hare-brained questions, however. Yes, we want to say that *the author of the page itself* must locally provide a label for each interactive widget, that is to say each object that accepts user input. Not rely on the SW somewhere and some day to fill in the gaps. And the knowledge that this element is such an interactive widget comes not from the element type but rather from an @role value. You see, we are trying to bootstrap into a more open system by writing formally in RDF. But for the critical near term, the vast majority of processors will not be RDF-breathing but will rather key off these @role values as specification-defined terms. Known to the programmer of the processor, and OK to implement in hard code. Another approach could perhaps be that we a) put the constraints in the OWL, because that's what our ontology is written in, and b) add a conformance clause to some profile spec that says to meet our profile you have to satisfy all the OWL properties with their constraints *locally* within the infoset provided by the current document (or maybe including documents linked by forward reference arcs, not backward arcs as in the RDF universe). Then the strictness would not come a_fortiori from the OWL semantics, but rather from the [e.g. XHTML] dialect and its strict use of @role. Call that plan B. Plan C: Write the population constraints in Schematron and publish a validation schema that runs in NVDL or some such technology. Do you have any suggestions as to how we can best do this? Or is it just to try on public-schemata-users? (not favorite answer) In any case, thank you very much for the help that you have already given. Al > >-Chris > >Al Gilman wrote: >> >> >>We in PFWG have a question [3] regarding the interaction between an OWL >>ontology [1] and the interpretation of document instances linked to this >>abstract model [2] via the XHTML @role attribute [4]. >> >>[1] Roles for Accessible Rich Internet Applications (WAI-ARIA Roles) -- >>Appendix A, Implementation >> >>latest public version: >>http://www.w3.org/TR/aria-role/#a_implementation >>Group draft (Member link): >>http://www.w3.org/WAI/PF/Group/GUI/#a_implementation >> >>[2] Roles for Accessible Rich Internet Applications (WAI-ARIA Roles) -- >>Building Accessible Applications >> >>latest public version: >>http://www.w3.org/TR/aria-role/#Building >>Group draft (Member link): >>http://www.w3.org/WAI/PF/Group/GUI/#Building >> >>I tried to ask on the semantic-web and public-rdf-in-xhtml-tf list, but >>have met with silence. >> >>[3] http://lists.w3.org/Archives/Public/semantic-web/2007Feb/0136.html >> >>I suppose I could try public-schemata-users, but what I actually want >>to do is to reach somebody who can and will help us. >> >>Can you advise me how I should address my request in order to get >>some support in understanding this issue and in applying OWL >>appropriately? >> >>Al >>/chair, PFWG >> >>[4] XHTML Role Attribute Module >> http://www.w3.org/TR/xhtml-role >> > >-- >Dr. Christopher A. Welty IBM Watson Research Center >+1.914.784.7055 19 Skyline Dr. >cawelty@gmail.com Hawthorne, NY 10532 >http://www.research.ibm.com/people/w/welty
Received on Wednesday, 14 March 2007 00:59:06 UTC