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

State of rich annotations

From: Bijan Parsia <bparsia@cs.man.ac.uk>
Date: Wed, 2 Jul 2008 17:27:32 +0100
Message-Id: <98BF75C0-5C8A-4584-AFC3-7B71E427F89C@cs.man.ac.uk>
To: OWL Working Group WG <public-owl-wg@w3.org>

Sorry, I meant this to go out earlier, but got caught. So it's a bit  

There are several dimensions of annotation support:

1) What can be annotated (entities (OWL1), axioms (OWL2 thus far),  
and annotations themselves (?))
	This also affects the "projection function", i.e., what objects are  
annotated. If a class doesn't have an explicit annotation, is it part  
of what's annotated? I.e., the empty annotation vs. the null  
annotation. This matters a lot for 2b.
	a) Should this be fixed or user definable?

2) The semantics of annotations, which can be organized in several  
	a) They can affect the object level entailments (OWL1,  
RichAnnProposal (RAP) mustUnderstand)
	b) They can have annotation entailments (not in OWL2, RAP) (e.g.,  
Dublin Core inheritance)
	c) The semantics of the annotations (OWL, logic programming,  
something else)
	d) Shared domain (punning), separate domain (OWL2 via no semantics;  
metaviews), or multiple domains (RAP)

3) The syntax of annotations
	a) Single assertions (OWL1) or blobs (RAP)
	b) Always rdf/owl, or arbitrary XML
	c) Implicit spaces only,or named spaces
		if named, each assertion must be explicitly placed in a space  
(RAP), or e.g., every annotation property P1 assertion goes into a  
designated space (MetaViews)
	d) how to deal with annotations on annotations, esp. in RDF syntax.

4) Fallback (behavior when you don't use any of the machinery).

Received on Wednesday, 2 July 2008 16:35:18 UTC

This archive was generated by hypermail 2.3.1 : Tuesday, 6 January 2015 21:42:05 UTC