[Pellet-users] SWRL Rules & Reasoning

[Michael Smith]

On Wed, 2008-07-02 at 13:31 +0200, Lamia Abo Zaid wrote:

> Excuse me but I’m really a beginner with SWRL rules. My question is
> how does pellet implement SWRL rules, does it  apply it to the
> asserted ontology only or does it first calculate the inferred types
> and then apply the rules?

Rules are evaluated as part of the iterative expansion process -- they
apply to both asserted and inferred facts.

> Can we change this order via code ?

No.  Rule application before or after the DL reasoning would not be
complete.

> When I call the classify() method of the reasoner does that evaluate
> the rules only ?

No, it evaluates all ontology axioms.

> how is this related to InferredOntologyGenerator () ?

That's part of the OWLAPI [1], not Pellet.  The Javadoc within the
OWLAPI says

        Generates an ontology based on inferred axioms which are
        essentially supplied by a reasoner. The generator can be
        configured with <code>InferredAxiomGenerator</code>s which
        generate specific kinds of axioms e.g. subclass axioms.
        
Given that, you could expect a properly configured
InferredOntologyGenerator to include results generated by your rules (if
such results can be inferred).


> Can you point me to any detailed documentation of what the methods in
> the pellet API do ?

Such documentation for end users doesn't exist.  There is some Javadoc
in the source (which is part of the distribution), but we generally
encourage people to use the OWLAPI or Jena [2] interfaces, which are
documented by the maintainers of those APIs.


-- 
Mike Smith

Clark & Parsia

[1] http://owlapi.sourceforge.net/
[2] http://jena.sourceforge.net/