W3C home > Mailing lists > Public > public-owl-wg@w3.org > November 2007

Re: ISSUE-67 (reification): REPORTED: use of reification in mapping rules is unwise

From: Ian Horrocks <ian.horrocks@comlab.ox.ac.uk>
Date: Mon, 26 Nov 2007 17:48:35 +0000
Message-Id: <4431EF62-1B2E-40C5-B72D-620262E0BC4F@comlab.ox.ac.uk>
Cc: Giorgos Stoilos <gstoil@image.ece.ntua.gr>, "'Boris Motik'" <boris.motik@comlab.ox.ac.uk>, "'OWL Working Group WG'" <public-owl-wg@w3.org>
To: Jeremy Carroll <jjc@hpl.hp.com>

On 22 Nov 2007, at 10:47, Jeremy Carroll wrote:

> Giorgos Stoilos wrote:
>> Hi,
>> As far as I understand annotations are already pretty weak  
>> semantically in
>> OWL 1.1 (actually they have no semantics), thus mapping them into a
>> structure with weak semantics, like reification, seems harmless to  
>> me.
> There is harm concerning:
> a) backward compatibility ISSUE-72

How so? It will still be the case that OWL 1.0 DL ontologies will be  
valid OWL 1.1 DL ontologies. Of course there may be OWL 1.1 DL  
ontologies that are not valid OWL 1.0 DL ontologies, but this is  
always going to be the case.

> and
> b) OWL Full/OWL DL semantic compatibility ISSUE-63

As I mentioned in my earlier email, they are not completely  
compatibly now. Be that as it may, it seems to be a reasonable goal,  
and is enshrined in our charter, that "The existing compatibility  
between OWL DL and OWL Full should be preserved, and should be  
extended to new features wherever possible." This does not, however,  
preclude the inclusion of features for which such an extension proves  
difficult or impossible.

> ===
> OWL 1.0 Full treats annotations on entities in the same way as  
> other properties. It is likely to be very confusing if OWL 1.1 has  
> annotations on axioms with different semantics from annotations on  
> entities. Thus, in OWL Full, annotations on axioms should have the  
> semantics of other properties. Building such a semantics on top of  
> reification is known not to work.

In fact the semantics of annotations in OWL DL and OWL full are  
already different in practice: whatever the spec might say, all OWL  
DL tools/systems that I am aware of treat annotations as being  
completely void of semantics (and I never heard of any users  
complaining about this). This seems to suggest that the existing  
design (where annotations have weak semantics) does not meet the  
expectations/requirements of either implementors or users.


> Jeremy
Received on Monday, 26 November 2007 17:48:44 UTC

This archive was generated by hypermail 2.3.1 : Tuesday, 6 January 2015 21:42:00 UTC