RE: Suggestion to change RDF encoding of HasKey axioms [RE: I've implemented the change to the syntax of HasKey]

>-----Original Message-----
>From: Ivan Herman [mailto:ivan@w3.org]
>Sent: Saturday, March 21, 2009 7:41 AM
>To: Michael Schneider
>Cc: W3C OWL Working Group; Boris Motik
>Subject: Re: Suggestion to change RDF encoding of HasKey axioms [RE:
>I've implemented the change to the syntax of HasKey]
>
>If I put an RDF+RDF base semantics user's hat on, isn't it correct that
>the same property may be used both as an object and a data property? 

Yes.

(And if we /would/ allow for object/data property punning in DL,
the situation would be similar there, too.)

>Ie,
>in many cases I would not really really care about this distinction. 

You may always put every property in the owl:keyObjectProperties list 
and leave the data property list empty.

>For
>such cases isn't it correct that the separation of the key properties
>would look fairly artificial?

If you do not want the fine grained distinction, then maybe. 
But equally well it could be asked

* why the additional class?
* why more than one property?
* Why keys at all, we have IFDPs in Full?

Key axioms are pretty flexible beasts which give authors a lot of control
over how they want to define keys. And in Full, as much as possible of this
flexibility should be reflected in the RDF-Based Semantics.

I wouldn't call Keys a "typical" OWL Full language feature. (They are
special enough, that they deserve their own semantic condition table. :-))
But they are in OWL 2 now, so their intended semantics should be supported
as precisely as possible. Doing so, that's the job of the one with the 
RDF Based Semantics Editor's hat on. :-)

With the data property list, the included properties can then be
entailed to be data properties, without saying so. Currently,
the semantics only say that all the properties involved are, well,
properties.
If a property is intended to be a data property, this information has to
come from a different source (e.g. from an explicit declaration triple).

In your other mail:

>Sorry to have forgotten to add this to my previous mail. Again in OWL 
>Full I could also use annotation properties in a key, right? 

Yes, you can put everything you want into a key axiom (even cats and dogs,
as Peter likes to say :)). But a key axiom will simply ignore that 
something is (also) an annotation property. It will, however, infer
that something is (also) a data property, if you put it in the
data property list. And this additional information may or may not
be a win for reasoners or other processing tools, or simply for people
working with the ontology.

>Of course, 
>annotation properties are just (object) properties in owl full, but the 
>explicit requirement in the key axiom would still be a bit disturbing 
>for owl full users...
>
>Ivan 

Keys are for those who need/desire them. No OWL Full user is forced to 
use key axioms in his ontology. But if an OWL Full user uses them,
and provided that he knows well what they are for, then I do not see
a reason to bother about the explicit distinction between data and
object properties.

Cheers,
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
=======================================================================