W3C home > Mailing lists > Public > public-owl-wg@w3.org > September 2008

Action 211: Review of Conformance Document

From: Bernardo Cuenca Grau <Bernardo.Cuenca.Grau@comlab.ox.ac.uk>
Date: Mon, 22 Sep 2008 11:51:20 +0100
Message-ID: <48D778A8.7030402@comlab.ox.ac.uk>
To: OWL Working Group WG <public-owl-wg@w3.org>

Hi,

This email addresses my action concerning the review of the conformance 
document. I have considered the latest version, which is integrated with the
test document.

http://www.w3.org/2007/OWL/wiki/Test#Conformance_.28Normative.29

Cheers,

Bernardo

------

**Beginning of Section 2**

I think the section is missing an introduction that explains what is the 
purpose and scope of the conformance definitions. Otherwise the content 
of the document can be confusing (see below).

** Section 2.1.1 **

I agree with Mike Smith's and Michael Schneider's comments about the 
document only mentioning the RDF syntax.
For example, if we have an OWL 2 EL ontology in functional
syntax that complies with the OWL 2 EL grammar as defined in the 
profiles document, is it then an OWL 2 EL document? The role of the RDF 
parsing process mentioned in the definition
seems to me unclear in this case. It seems that the syntactic 
conformance definitions are only focused on and targeted towards the RDF 
syntax; however, the scope of these definitions
should be clearly stated in the document to avoid confusion. I suggest 
two possible fixes:

1) To make it clear in the section that the syntactic conformance 
definition is only intended for ontologies in RDF
2) To change the definition of syntactic conformance to include all 
possible syntaxes

**Section 2.2.1**

The definitions of semantic conformance should be slightly rephrased. 
For example, the document says:

``*An OWL 2 DL entailment checker* is an OWL 2 entailment checker that 
takes OWL 2 DL ontology documents as input, and uses the Model Theoretic 
Semantics [OWL 2 Semantics <#ref-owl-2-semantics>]. It /MUST/ return 
True only when /O_1 / entails /O_2 /, and it /MUST/ return False only 
when /O_1 / does not entail /O_2 /. It /SHOULD NOT/ return Unknown. ''

Should be rephrased in the following way:

``*An OWL 2 DL entailment checker* is an OWL 2 entailment checker that 
takes OWL 2 DL ontology documents as input.  It /MUST/ return True only 
when /O_1 / entails /O_2 / with respect to the Model Theoretic Semantics 
[OWL 2 Semantics], and it /MUST/ return False only when /O_1 / does not 
entail /O_2 / with respect to the Model Theoretic Semantics. It /SHOULD 
NOT/ return Unknown. ''

Similarly with the other definitions.

** Definition of the OWL 2 RL Conformance **

 From the definition in the document it seems that a sound and complete 
OWL 2 DL reasoner may not be an OWL 2 RL entailment checker. I find this 
quite strange. According to the profiles document, OWL 2 RL can be seen 
as a syntactic fragment
of OWL 2 (there is a proper grammar that defines such fragment). I think 
that the problem is again that the scope of the definitions is not clear 
in the document, which causes ambiguity. In the current situation, the
profiles and conformance documents don't seem to be in sync with each other.
I suggest the following solution. We should have two notions of 
conformance for OWL 2 RL:

 1) OWL 2 RL Entailment Checker:  an OWL 2 entailment checker that takes 
OWL 2 RL ontology documents  as input, and uses the Model Theoretic 
Semantics [OWL 2 Semantics <#ref-owl-2-semantics>]. It /MUST/ return 
True only when /O_1 / entails /O_2 /, and it /MUST/ return False only 
when /O_1 / does not entail /O_2 /. It /SHOULD NOT/ return Unknown.
 2) OWL 2 RL RDF Entailment Checker: defined as in the current 
conformance document

Then we should show that there is a correspondence for OWL 2 RL 
ontologies (as in Theorem 1 in the profiles document).
Received on Monday, 22 September 2008 10:52:10 UTC

This archive was generated by hypermail 2.3.1 : Tuesday, 6 January 2015 21:42:07 UTC