W3C home > Mailing lists > Public > public-owl-wg@w3.org > April 2009

Re: Datatype map conformance designation needed

From: Ian Horrocks <ian.horrocks@comlab.ox.ac.uk>
Date: Tue, 7 Apr 2009 18:16:15 +0100
Message-Id: <4B2C1BEE-1574-4B87-B44C-81CFE7F3B80C@comlab.ox.ac.uk>
Cc: W3C OWL Working Group <public-owl-wg@w3.org>
To: Jonathan Rees <jar@creativecommons.org>
Jonathan,

There may be some misunderstanding here. An "ontology document that  
requires a particular datatype map that extends the OWL 2 datatype  
map" is not an OWL 2 DL Ontology Document according to the  
specification. Perhaps this wasn't clear from the existing wording in  
conformance, so I added text to make it explicit (see [1]).

The current conformance conditions do not specify how tools should  
behave when presented with non-conformant documents. It should  
hopefully be clear to users and/or tool builders that consuming and/ 
or producing such non-conformant documents is "risky", and that it  
will compromise interoperability. The OWL spec deliberately tries to  
minimally constrain tool design, and explicitly specifying their  
behaviour in case they are presented with non-conformant documents  
seems to be an unnecessary constraint.

Hopefully this is sufficient to address your concerns.

Regards,
Ian

[1] http://www.w3.org/2007/OWL/wiki/index.php? 
title=Conformance_and_Test_Cases&diff=21320&oldid=21268



On 1 Apr 2009, at 18:45, Jonathan Rees wrote:

> Synopsis: To promote interoperability, the conformance document should
> define a designation to apply to documents that conform to the OWL 2
> datatype map, as opposed to some extended map.
>
> Consider an OWL 2 ontology document that requires a particular
> datatype map that extends the OWL 2 datatype map.  This document
> cannot be given to an OWL 2 processor that does not understand that
> datatype map, or else the document will not be understood correctly.
> For example, it may be found consistent when in fact it is  
> inconsistent,
> or vice versa.
>
> The conformance document lists two kinds of document conformance,
> syntactic conformance and datatype conformance.  If a document is
> described as having a particular kind of syntactic conformance, then
> users will be aware that problems are likely to result if the document
> is processed by a tool requiring a more stringent kind of conformance.
>
> Datatype conformance is not different from syntactic conformance in
> this regard.  It is just as important to be able to distinguish a
> document requiring an extended datatype map from one that only
> requires the OWL 2 datatype map, as it is to distinguish a document
> requiring OWL 2 Full conformance from one that only requires OWL 2 DL
> conformance.
>
> I would not presume to say exactly what the designation should be,
> but an adjective, adjectival phrase, or prepositional phrase would be
> best so that conformance to the OWL 2 datatype map can be expressed
> succinctly in conjunction with syntactic document conformance.  For
> example, if the designation were the adjective "datatype-conformant"
> (just a for-instance, not a term I advocate), one could say that a
> document is a "datatype-conformant OWL 2 DL ontology document".
>
> Best
> Jonathan
>
> (You may also consider this message to be in the way of a review of  
> Conformance and Test Cases.  The suggestion is similar to the  
> "strict" mode idea in Alan's review [1]. The issue connects to AR1  
> [2] but is a more general question concerning how users relate to  
> datatype map extensions.)
>
> [1] http://www.w3.org/mid/ 
> 29af5e2d0903312316y60879c74yfb9ce5e23998699@mail.gmail.com
> [2] http://lists.w3.org/Archives/Public/public-owl-wg/2009Feb/ 
> 0272.html
>
>
Received on Tuesday, 7 April 2009 17:16:54 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Tuesday, 7 April 2009 17:16:54 GMT