- From: Sandro Hawke <sandro@w3.org>
- Date: Mon, 13 Nov 2006 10:06:05 -0500
- To: Dave Reynolds <der@hplb.hpl.hp.com>
- Cc: public-rif-wg@w3.org
> Sandro Hawke wrote:
> > Dave Reynolds wrote:
> >> Francis McCabe wrote:
> >>> How about Rule Ontology?
> >> Of course, though that's presumably already in progress with RIFRAF.
> >
> > There are two separate things, here, right?
> >
> > 1) An ontology of rule systems and rule languages
> >
> > Instances: each different rule system (Jess, Prova, Blaze
> > Advisor, JenaRules, etc), as on
> >
> > http://www.w3.org/2005/rules/wg/wiki/List_of_Rule_Systems
> > and in the answers to the questionnaire.
> >
> > The WG will probably provide instance data for a dozen or two, of
> > the hundreds that probably exist.
> >
> > This is what I think RIFRAF is about.
> >
> > 2) An ontology of rules and rule sets.
> >
> > Instances: each different rule and/or rule set
> >
> > Millions of these exists; users of RIF could be imagined as
> > authoring instance data in this ontology. I often think about it
> > that way (I think all syntaxes are just ways of serializing
> > triples), but I'm agnostic on whether the rest of the WG thinks
> > about it this way.
> >
> > I think Frank was talking about #2 when he said "rule ontology".
>
> I agree these are two separate things but isn't (at least part of) the
> intention of the RIFRAF work to inform #2 by identifying common shared
> concepts? Perhaps I should have said "in progress via RIFRAF".
> Personally I'm not particularly motivated by #1 but had thought it had
> been justified as input to #2.
>
> I was assuming the RIFRAF work would generate outputs such as "lots of
> rule language seem to have feature X, how will we handle feature X?" and
> feature X is then a shared concept in the sense of #2.
>
> For example:
> "ah, there is a cluster of languages here that support what we might
> call 'action rules' which invoke some sort of action in their
> consequent, so we need to make sure the #2 ontology includes concepts
> like :Action, :ActionRule and an extensible library of specific types of
> action. Here's the specific sorts of actions that seem to be needed ..."
Yes, absolutely, that's my understanding as well.
-- Sandro
Received on Monday, 13 November 2006 15:06:30 UTC