Re: Declarative limitations of OWL/DAML

lchewhun@dso.org.sg wrote:
> 
> I'm interested in what CANNOT be represented (declared) in OWL/DAML. Your
> comments or pointers to papers are much appreciated.
> 
> What I'm hoping to do is to be able to represent constraints (or axioms or
> rules) in a DAML ontology so that an inferencing engine can check the ontology
> for consistency.  

Ian Horrocks and Benjamin Grosof have some working papers such as [1] which
compare expressivity of DAML/OWL to rule or logic programming languages. In
particular, they point out that DL languages can't represent predicates
involving multiple variables or involving predicate chains other than transitive
closures. Thus a constraint such as "my grandfather must be older than me" would
be tricky on both counts.

Personally I think the semantic web will need a rich constraint expression
language eventually, perhaps it could build upon something like OCL by using
path expressions to identifying the values being constrained?

Dave

[1] http://www.cs.man.ac.uk/~horrocks/Private/DLP/DLP-v9-brief.pdf

-- 
Hewlett-Packard Laboratories    | Phone: +44-117-3128165
Filton Road, Stoke Gifford      | FAX:   +44-117-3128925
Bristol BS34 8QZ, UK            | dave.reynolds@hpl.hp.com

Received on Tuesday, 25 March 2003 06:01:54 UTC