- From: Christine Golbreich <cgolbrei@gmail.com>
- Date: Tue, 4 Nov 2008 23:59:06 +0100
- To: ivan@w3.org
- Cc: public-owl-wg@w3.org
Hi Ivan Thank you very much for your review of the Requirements/OWL2 New Features at [1]. Below I commented the revisions I have done to address your comments. These revisions are included in the updated version of the document at [1]. Please let me know how you find them. Thanks Christine [1] http://www.w3.org/2007/OWL/wiki/OWL2NewFeatures 2008/11/3 Ivan Herman <ivan@w3.org>: > Christine, > > I have very small comments only, all editorial > > - 2.2.2: although the formal spec makes it clear that the, say, > MinCardinality construct in the functional spec encompasses both the new > usage and the OWL 1 restrictions (ClassExpression being optional) it may > be good to emphasize that for OWL 1 readers. emphasized 'both' in italic in the Feature description: "both qualified or unqualified cardinality restrictions are possible" > This is all the more > important because in the RDF mapping there is actually a fork, and the > cardinality with an explicit class is mapped on a different term than > the one without... Is there an example in the Use Cases that could be > added that does _not_ use a class? That could be added, for example. added an example with an explicit comment > - 2.2.5: I wonder whether it is possible to give a short (though not > necessarily complete) description of the global restrictions on how to > use property chains. The current text refers to the syntax document > which is, well, not really easy to read and certainly on the level of > the readers of this document.... Agree, but it's rather challenging : global restrictions are not trivial ... and as said in the syntax, it's rather technical, it's why I refered to it :-) I'll see whether I might add something simple, and still correct without going into too much details, but it's not sure. > - 2.2.6: AFAIK, and the functional syntax seems to say that, a key can > be both a datatype and an object property. However, the text suggests > otherwise: "However, keys, aka inverse functional datatype properties" > which might lead to the user that only datatype properties can be used > in this position. OK, revised it > I also wonder whether it is possible to find an example where more than > one properties appear in the HasKey clause. The fact that one can define > an instance through the combination of properties is very powerful and > would deserve more emphasis there... added > - 2.3.1: I think it is worth emphasizing that we also have owl:real as a > special datatype, as well as rdf:text. The current text suggests that we > use 'most XML Schema datatypes' which suggests that we use a subset of > those. Which is true but we also have some others... revised this section > - 2.3.2: just flagging this; afaik, we do not have a decision on the > inclusion of n-ary datatypes yet... added an Editor Note > - 2.5.1: it might be good to have examples for the (new) possibilities > of defining ranges and domains for annotations and defining sub > properties for annotations. I am not sure any of the use cases have > that, though... I'm waiting for their inclusion in the syntax doc, to give an example in the right syntax. > - It may be worth adding the top and bottom object properties to > features. Boris had some great example for the type of expression one > might have with those around. I don't remember Boris' examples. But an example from a real UC might be the top property used for the RO ontology (ontology of relations) as a root of the relations ? However, as some members said at the F2F, their use is a little dangerous. So do we really want to stress these properties in adding a new feature for them ? > I hope this helps! > > Thanks a lot > > Ivan > > Christine Golbreich wrote: >> A revised version of the Requirement is now available at >> http://www.w3.org/2007/OWL/wiki/OWL2NewFeatures. >> >> It is based on the discussions & strawpolls at the F2F and review comments. >> The objective of this updated version "New Features & Rationale", is >> to focus on an overview of the new features that have been added to >> OWL 2 and of their rationale. Find below the main changes: >> >> - renamed "Requirements" into "New Features & Rationale" >> - moved the core section "Features" up, to be the first, as unanimously agreed >> - rolled sections 4 (req) into 5 (features) >> - moved the use cases to an appendix, as wished by the majority (strawpoll) >> - refered to the UCs from the section "features" >> - cut section 2 (users and applications), as proposed mainly by Bijan >> - completed and updated some features >> - All the UCs in the Appendix follow a common pattern as asked >> Elis'as review earlier: Overview, Features, Example, Literature. >> - separated References and UCs Biblio >> >> There might still be some polishing to do (add links, etc. ) but it >> gives a good idea of the new proposed version. >> It would be helpful now to check whether all content is up-to-date >> w.r.t last decisions in particular at the F2F or wherther the document >> should be fixed in some places. >> >> All your comments and feedback are welcome. >> >> Christine
Received on Tuesday, 4 November 2008 22:59:54 UTC