- From: Dan Connolly <connolly@w3.org>
- Date: 31 Oct 2002 13:33:21 -0600
- To: "Peter F. "Patel-Schneider <pfps@research.bell-labs.com>
- Cc: Jim Hendler <hendler@cs.umd.edu>, Guus Schreiber <schreiber@swi.psy.uva.nl>, Dan Connolly <connolly@w3.org>, www-archive@w3.org, Massimo Marchiori <massimo@w3.org>
On Thu, 2002-10-31 at 12:40, Peter F. Patel-Schneider wrote: > > Given that we didn't get around to the semantics document, Yeah, that's kinda unfortunate... I guess I hijacked the agenda in some ways... Meanwhile, after the telcon, JimH and Guus and I talked about stuff to do about this rather than just letting it sit for a week. We've got reviewers in mind and stuff. Expect to hear from JimH. [...] > Points with respect to the New Semantics Document Er... is there some reason you didn't include a pointer? > - Move forward to make this a WG document? I hope we'll be prepared to decide next week. Certainly by the following week. > - Organization > - Introduction > - Abstract Syntax > - Direct Model-Theoretic Semantics > - Mapping to RDF Syntax > - RDFS-Compatible Model-Theoretic Semantics > - Commonalities, Fast OWL, Large OWL > - Correspondence between Abstract OWL and Fast OWL (Informative) > * What else is needed? I can't think of anything. > * What should be removed? I don't see anything. > - Stances on Open Issues > - 4.4 - Extra Logical Feature Set - only imports > - annotations - easy to handle in abstract syntax > - harder in RDF syntax (and not yet done) ok by me... > - 4.6 - equivalentTo - no equivalentTo anywhere > - sameIndividualAs carries very similar meaning in Large OWL > * I believe that this treatment is a basis for closing the issue yup; I'd prefer to just call it sameAs, since the domain/range isn't actually constrained to individuals/Things. > - 5.5 - Lists are handled directly, and in a minimal fashion > - I believe that it works in Fast OWL > * I believe that this treatment is a basis for closing the issue I'd like to study the details of this a bit more, but there's a good chance I won't get to it. I'm not opposed to publishing in any case, but I might not be ready to close this issue. > - 5.6 - imports - only partly handled > - easy to handle in abstract syntax, as in my email message > - harder in RDF syntax (and not yet done) That meets my expectations. > - 5.9 - malformed restrictions > - currently different from DAML+OIL > - a restriction with two onProperties causes the properties to > be identified (as if there was a samePropertyAs axiom) > - should this be changed back to the DAML+OIL situation > - a restriction with multiple onProperties or other properties > implies an equality between the satisfying sets of individuals > * can go either way, but the change from DAML+OIL requires a decision I don't see much difference either way. The guide should clearly say "don't do that." What happens when you do, I don't much care; whatever makes drafting the semantics easiest is OK by me. > - 5.19 - classes as instances > - allowed in Large OWL, forbidden in Fast OWL > * this is in accordance with the direction agreed on at the F2F yup, ok. > - 5.7 - no datatype range constructs That's ok by me, but I think DAML users expect this functionality, and we need to check that they find this acceptable, for now. We may need to mark this POSTPONED in the issues list, rather than claiming to have addressed it. > - 5.8 - datatypes handled in a similar manner to current RDF proposal whatever that means. 1/2 ;-) > - 5.13 - internet media type - no stance ok by me. > - 5.14 - ontology versioning - no solution ok by me. > - 5.17 - XML presentation syntax - not relevant right. > - 5.18 - unique names assumption - no solution ok by me, though I'm on the hook to propose the Distinct/AllDifferent thingy. > > WARNING: I will be making proposals based on this document. I look forward to them! -- Dan Connolly, W3C http://www.w3.org/People/Connolly/
Received on Thursday, 31 October 2002 14:33:07 UTC