RE: Progress on QRG

>-----Original Message-----	
>From: public-owl-wg-request@w3.org [mailto:public-owl-wg-request@w3.org]
>On Behalf Of Jie Bao
>Sent: Saturday, May 02, 2009 6:45 AM
>To: public-owl-wg
>Cc: Peter F. Patel-Schneider; Elisa Kendall; Deborah L. McGuinness
>Subject: Progress on QRG
>
>Hi All
>
>I made a few changes to QRG. The diff to the version sent before the
>Apr 29 telcon [1] is in [2]

[...]

>To be discussed

[...]

>* Whether owl:distinctMembers is deprecated? 

No, there was never a formal vote on its deprecation, although I would have preferred to see it deprecated. It really shouldn't be used in new ontologies anymore, IMO, because it has been obsoleted by the more general term "owl:members".

>It is not used in Syntax,

Obviously not.

>nor used in the mapping from Syntax to RDF syntax (thus, effectively
>an OWL 2 editor will not produce a RDF document using
>owl:distinctMembers). 

Yes, it wouldn't make much sense to have two different mappings to RDF for the same OWL 2 construct in an ontology, even if the results are semantically identical. 

>It is indeed used (and only used) in parsing an
>OWL 2 RDF syntax document into the functional syntax ([4], Table 16),
>for clearly backwards compatibility to OWL 1. 

Yes.

>I'm not sure about its status. The current set of documents 
>is quite silent on that, which may confuse some users. 

It is also treated in the RDF-Based Semantics, of course, for the same reason that it is still used in the reverse mapping.

>Note that I'm not arguing for its deprecation,
>I'm asking the right way to document it.

I think we should be very careful with the use of the term "deprecation" in our documents. Deprecation of a feature means *at least* that people should refrain from using the feature in new ontologies. This is quite the opposite of what it normally intended for a feature that is listed in a Recommendation. So deprecation is, for me, a design change.

I am pretty ok, however, to follow Peter's suggestion to use another term such as "compatibility vocabulary" or whatever.

But I would be even more fine with dropping this table completely from QRG, and leave it to the core specifications to say what is deprecated or only supported for compatibility reasons and what not. For example, the RDF-Based Semantics states in Section 3 and Section 9 that there is one deprecated vocabulary term: owl:DataRange, for which we had a vote on its deprecation. And the RDF Mapping has these "compatibility tables", which at least implicitly make an assertion about which terms are "cool" and which are not.

>* In the Syntax, there is no mentioning of ontology properties. All
>the three built-in "ontology properties" are actually defined as
>annotation properties. However, in the RDF semantics, there are
>ontology properties. So it is to be discussed whether we should call
>them "Ontology Properties" or Annotation Properties for Ontologies".

I consider ontology properties to be primarily a feature of the RDF-Based Semantics. I don't see any real relevance for any other part of OWL (2). Ontology properties are properties that relate two ontologies in some way, and in OWL (2) Full one can infer from a triple

  :x :someOntologyProperty :y

that :x and :y are ontologies.

What's more important, the term "owl:OntologyProperty" denotes a "part" of the OWL universe, which is related to another part denoted by owl:Ontology, so it has its role in the basic semantic structure of the RDF-Based Semantics. Further, the document refers to ontology properties in several places, so it's far from being "dead", at least in the RDF-Based Semantics. Trying to get rid of ontology properties would lead to non-trivial changes to the RDF-Based Semantics. 

Unlike owl:DataRange (and owl:distinctMembers), there is no such clear replacement for ontology properties. In particular, it is not true that every ontology property in the RDF-Based Semantics is an annotation property elsewhere in the spec: owl:imports is an ontology property, but certainly not an annotation property. Having two names for essentially the same concept, as it is true for owl:DataRange/rdfs:Datatype and owl:distinctMembers/owl:members, is really confusing, so it's fine by me to discourage the use of one term for each such term pair. But "owl:OntologyProperty" isn't simply a synonym for "owl:AnnotationProperty", so the situation is different.

>** do we ever have a formal solution for owl:OntologyProperty?

No. And unlike for the deprecation vocabulary, owl:DataRange and owl:distinctMembers, there hasn't even ever been any discussion on owl:OntologyProperty, to my best knowledge. And given that our documents, specifically the RDF-Based Semantics, are now in LC, it is far too late to start such a discussion. 

And, as far as I am concerned, I would be very much against the deprecation of owl:OntologyProperty.

Best,
Michael

--
Dipl.-Inform. Michael Schneider
Research Scientist, Dept. Information Process Engineering (IPE)
Tel  : +49-721-9654-726
Fax  : +49-721-9654-727
Email: michael.schneider@fzi.de
WWW  : http://www.fzi.de/michael.schneider
=======================================================================
FZI Forschungszentrum Informatik an der Universität Karlsruhe
Haid-und-Neu-Str. 10-14, D-76131 Karlsruhe
Tel.: +49-721-9654-0, Fax: +49-721-9654-959
Stiftung des bürgerlichen Rechts, Az 14-0563.1, RP Karlsruhe
Vorstand: Prof. Dr.-Ing. Rüdiger Dillmann, Dipl. Wi.-Ing. Michael Flor,
Prof. Dr. Dr. h.c. Wolffried Stucky, Prof. Dr. Rudi Studer
Vorsitzender des Kuratoriums: Ministerialdirigent Günther Leßnerkraus
=======================================================================