W3C home > Mailing lists > Public > public-owl-dev@w3.org > April to June 2007

Re: RDFS compatibility information in OWL-DL documents

From: Bijan Parsia <bparsia@cs.man.ac.uk>
Date: Tue, 17 Apr 2007 17:17:28 +0100
Message-Id: <BBDA6EA8-C7BE-4A52-9234-25E16A7ACAF7@cs.man.ac.uk>
Cc: public-owl-dev@w3.org
To: Alan Ruttenberg <alanruttenberg@gmail.com>

On Apr 15, 2007, at 11:49 AM, Alan Ruttenberg wrote:

> It is often the case that someone using OWL wishes to target some  
> information for specifically for tools which understand RDF/RDFS in  
> some way. A good example would be wanting to indicate to particular  
> property should be considered a user interface label for a rdf  
> "follow your nose" type of browser such as the tabulator (http:// 
> www.w3.org/2005/ajar/tab).
> In this case, the way to communicate this information to the  
> tabulator is to make the property in question a subproperty of  
> rdfs:label. However, doing such a thing brings an ontology in to  
> OWL-Full.
> Another example would be including the rdfs version of the OWL  
> definitions (owl:AnnotationProperty  rdfs:subPropertyOf  
> rdf:Property) for those tools where it might be useful.
> I am wondering whether we can add some standard way of indicating  
> such information in an OWL-DL document, either syntactically, or  
> via an appropriately named annotation property that we can  
> advertise as being for this use.

Or just a better profile.

We could get the effect of OWL 1.1 entity annotations (or even  
punning) by defining an ad hoc species "OWL DL with enriched  
annotation properties) to treat as if it were in a separate  
document". We'd certainly update Pellet to support that (it's a very  
small change overall).

The other way is to keep the annotations and the domain model in  
separate documents, then have a third document which imports them.  
There are some advantages to keeping the presentation information  
separate, though this particular way of doing it is can be clunky.

Hmm. You could annotate properties that were for display or were  
labels. The main problem with that is that existing tools wouldn't be  
sensitive to it. Fresnel could be another solution since it does  
separate the presentation fairly neatly.

Really, I think the smart thing is to figure out what would work  
reasonably for users and then just get everyone to go with that  
convention. The infrastructure is small enough that it's feasible,  
if...annoying :)

Received on Tuesday, 17 April 2007 16:17:55 UTC

This archive was generated by hypermail 2.3.1 : Tuesday, 6 January 2015 20:58:15 UTC