Re: question about annotations

From: Jeremy Carroll <jjc@hpl.hp.com>
Subject: question about annotations
Date: Thu, 24 Jan 2008 09:16:39 +0000

> During the discussion yesterday I got a bit confused about say the 
> following set of triples in OWL 1.1
> 
> eg:c rdf:type owl:Class .
> eg:c eg:dp "A string" .
> 
> =====
> 
> If I put these two triples through an OWL 1.0 'make it DL patch-up' 
> program, I would get
> 
> eg:dp rdf:type owl:AnnotationProperty .

Maybe.  I don't know if there are tools that do this patch.

> If I put these through an OWL 1.1 (with punning) 'make it DL patch-up' 
> maybe I should get
> 
> eg:dp owl:DataProperty .

Maybe again.   This would be more likely, I think.

> with eg:c being punned as a class and an individual.


> This would suggest that no annotations are needed at all in OWL 1.1 - 
> and then I get confused because we have discussions about annotations.

Well, this depends on what you think annotations are supposed to be.   

One view is that annotations are extra-logical fluff added to
ontologies, and thus are still needed.  In this view the fact that
annotations were turned into facts in OWL Full was a very unfortunate
consequence of the "triples uber alles" RDF view of life, the universe,
and everything.  Removing the connection between annotations and facts
removes a silly aspect of OWL that only gets in the way.

Another view is that the major use for annotations is to add (logical)
information to classes.  In this view it would be natural to remove
annotation properties because they were only added to allow this sort of
information to (uneasily) exist in OWL DL.

> =====
> 
> Please clarify, are there annotations on entities in OWL 1.1, and if so, 
> why?

Yes, and see above.

> Jeremy

peter

Received on Thursday, 24 January 2008 09:36:10 UTC