LC Comment on Entailment Regimes: Insufficient Definition of "Entailment Regime"

Dear all!

Document: SPARQL 1.1 Entailment Regimes
State: LCWD

In Section 1.3, condition 1, the specification of "entailment regimes" 
states:

     1. A subset of RDF graphs called well-formed
        for the regime

I believe that this condition is not sufficient for all entailment 
regimes in the document.

Rational:
---------

Take, for example, the following queried graph O and the (ground) BGP O1:

     O = {
         :p rdf:type owl:TransitiveProperty .
     }

     O1 = {
         :p rdf:type owl:AsymmetricObjectProperty .
     }

Under the OWL 2 DL syntax [1], both O and O1 are syntactically valid 
ontologies (in RDF graph form), so they meet condition 1 for the OWL 2 
Direct Semantics Entailment Regime. However, the entailment query "(O, 
O1)" is /not/ valid, as the combination of both graphs hurts the global 
restrictions in Chap. 11 of [1]: transitive properties must not be 
asymmetric properties. This is explained at the end of Chap. 3 of [2]:

     For ontology equivalence to be decidable,
     O1 needs to satisfy this restriction w.r.t.
     O and vice versa.

(I believe that this condition of the Direct Semantics should not be 
restricted to /equivalence/, but is also needed for entailment).

So, I believe that a entailment query "(O, O1)" has to be disallowed for 
the OWL 2 Direct Entailment Regime, at least if this regime is 
restricted to the OWL 2 DL syntactic fragment of OWL 2.

Proposal:
---------

Instead, I suggest, to change the condition as follows:

     1. A set of /pairs of RDF graphs/ called well-formed
        for the regime

In addition, in the tables for all entailment regimes, the row "Legal 
Graphs" should be replaced by "Legal Graph Pairs" or something like 
"Legal Entailment Queries", with the corresponding texts being updated.

Final Note:
-----------

If, against my current understanding, the OWL 2 Direct Entailment Regime 
is meant to be applicable to the full OWL 2 Structural Specification 
[1], and not only to the OWL 2 DL fragment with the global constraints 
to retain decidability (and other restrictions), then my comment is 
moot. But then, the document should be carefully checked that there is 
really no dependency on the OWL 2 DL fragment anywhere else.

[1] http://www.w3.org/TR/2009/REC-owl2-syntax-20091027/
[2] http://www.w3.org/TR/2009/REC-owl2-direct-semantics-20091027/

Best regards,
Michael

-- 
Dipl.-Inform. Michael Schneider
Research Scientist, 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
Stiftung Az: 14-0563.1 Regierungspräsidium Karlsruhe
Vorstand: Dipl. Wi.-Ing. Michael Flor, Prof. Dr. rer. nat. Ralf Reussner,
Prof. Dr. rer. nat. Dr. h.c. Wolffried Stucky, Prof. Dr. rer. nat. Rudi 
Studer
Vorsitzender des Kuratoriums: Ministerialdirigent Günther Leßnerkraus
==============================================================================

Received on Tuesday, 26 July 2011 12:04:13 UTC