New OWL 1.1 draft documents, some questions

Dear all,

I just had a look at some of the new OWL 1.1 draft documents posted at 
[1] and have a few questions (and remarks).

The first is about the notion of 'complex properties' used in the 
Overview. It seems that the expressiveness added by SROIQ semantics 
involves non-complex properties, but it does not really say anywhere 
what these non-complex properties are and what distinguishes them from 
their complex counterparts. My intuition is that complex properties are 
those defined by complex role inclusion axioms, but I couldn't find any 
statement that confirms or refutes this.

My other questions concern the OWL XML syntax document (is this the 
right forum?). It states that "all URIs in an ontology are assuming to 
be relative to the xml:base of the ontology document". Apart from the 
typo, I wonder whether this requirement comes down to restricting the 
values of the owl:URI attributes to *relative* URI's only (since the 
examples only use  fragment identifiers).

And secondly, how is the relation between the xml:base and 
owl:ontologyURI attributes to be interpreted? Obviously the ontologyURI 
URI does not comply with the just mentioned requirement. Does the 
standard allow for definitions in the same ontology to be stated in two 
different XML files with differing xml:base values? As owl:URI attribute 
values are relative to the xml:base, this results in two namespaces 
within the same ontology.

A further complication, in my view, is the fact that the standard makes 
no distinction between identifiers of and references to owl entities. 
Intuitively, an XML serialization of an ontology would introduce 
entities using declarations (providing them with identifiers), and then 
refer to these entities in the rest of the serialization (i.e. the 
axioms). This might well be too RDF-ish, but the current situation seems 
a bit inconsistent.

Wouldn't it be better to use the owl:URI attributes to refer to elements 
of the ontology, i.e. the uri's are resolved to the owl:ontologyURI, or 
*any* other ontology out there, independent of the xml:base. 
Users/agents can then use regular xml:id's to identify the separate 
statements about elements of the ontology in the XML serialization. 
These xml:id's are then of course resolved to the xml:base of the 
document they reside in.


PS There's another typo in the sentence about typos.


Drs. Rinke Hoekstra

Email:   Skype:  rinkehoekstra
Phone: +31-20-5253499    Fax:   +31-20-5253495

Leibniz Center for Law,         Faculty of Law
University of Amsterdam,           PO Box 1030
1000 BA  Amsterdam,            The Netherlands

Received on Thursday, 26 October 2006 16:50:19 UTC