Re: proposed response to OWL LC documents

As an afterthought:

Perhaps we should include a fourth point about the fact that 
RDF/XML is not any more the only normative exchange syntax. This 
may very well hamper interoperability between the SKOS/RDF and 
OWL2 world.

Guus


Guus Schreiber wrote:
> This is a *proposed* response. Pls comment.
> Guus
> 
> ------------------------
> 
> Dear OWL WG,
> 
> The Semantic Web Deployment (SWD) Working Group has reviewed the OWL2 
> Last Call documents. We apologise that these comments are provided to 
> you after the deadline. We hope you will still be able to consider them.
> 
> NOTE: Our comments are given from the perspective of the work of SWD on 
> SKOS [1], as SKOS is based on RDF/OWL.
> 
> 
> 1. The SWD WG notes that some of the extensions provided by OWL2 appear 
> to be useful for SKOS. For example, property disjointness can be used in 
> specifying the semantics for SKOS mapping relations. Although currently 
> not needed in the SKOS semantics, we can foresee use cases for new 
> property characteristics such as (a)reflexivity and asymmetry, e.g. for 
> specifying application-specific specializations of SKOS semantic 
> relations.  For alignment between SKOS and OWL DL the possibility to 
> define axioms about annotation properties is perceived as useful.
> 
> 
> 2. The SWD WG is disappointed about the way the OWL2 material is 
> presented, in particular the lack of using either an RDF/XML or an RDF 
> triple representation of OWL2. Even the " New Features and Rationale" 
> document [2] refrains from using such syntax. This makes the OWL2 
> documents inaccessible for the typical SKOS user. We request that the 
> OWL WG remedies this situation.
> 
> 
> 3. The SWD WG notes that most OWL2 documents give the impression that 
> OWL2 is just an extension of OWL1 DL, and not of OWL1 Full. For example, 
> the introduction of  OWL2 Direct Semantics document states [3]:
> 
>   [[
>     Since OWL 2 is an extension of OWL DL
>   ]]
> 
> Only one document clearly makes the OWL2 DL and OWL2 Full distinction 
> [4]. In our experience  OWL Full is the dominant OWL usage pattern for 
> SKOS. We therefore request that the OWL2 document are edited in such a 
> way that whenever the term "OWL2" is used, it is used to refer to the 
> OWL2 language as a whole (OWL2 DL and OWL2 Full). If OWL2 DL is meant, 
> it should be explicitly marked as such. We also request that the nature 
> and role of OWL2 Full are clearly stated in other central OWL documents, 
> not just in [4].
> 
> 
> We hope these comments are useful for you.
> 
> Best,
> Guus Schreiber
> on behalf of the Semantic Web Deployment Working Group
> 
> [1] http://www.w3.org/TR/skos-reference/
> [2] http://www.w3.org/TR/2008/WD-owl2-quick-reference-20081202/
> [3] http://www.w3.org/TR/2008/WD-owl2-semantics-20081202/#Introduction
> [5] http://www.w3.org/TR/2008/WD-owl2-rdf-based-semantics-20081202/
> 
>