W3C home > Mailing lists > Public > public-rif-wg@w3.org > January 2006

Re: RDF and OWL compatibility

From: Christopher Welty <welty@us.ibm.com>
Date: Mon, 9 Jan 2006 15:44:11 -0500
To: Michael Sintek <sintek@dfki.uni-kl.de>
Cc: public-rif-wg@w3.org, public-rif-wg-request@w3.org
Message-ID: <OF0A6561E5.B183E54E-ON852570F1.00702269-852570F1.0071E732@us.ibm.com>

public-rif-wg-request@w3.org wrote on 01/04/2006 02:35:50 PM:
> Jos de Bruijn wrote:
> > On Tue, 2006-01-03 at 18:21 +0000, Dave Reynolds wrote:
> >>Michael Sintek wrote:
> ...
> >>I agree that RDF compatibility requires support for quantification 
over RDF 
> >>properties. Whether this translates into a requirement for a 
higher-order 
> >>syntax for the rule language or into a requirement that the RDF 
mapping 
> >>should use some more straightforward "triple(s,p,o)" convention is a 
> >>separate decision.
> > 
> > 
> > I'm not sure whether support for quantification over RDF properties is
> > required, but I can imagine it could be.
> 
> The query <s,?p,o> was only one example where higher order would be
> needed for the "straightforward" mapping. 

Correction: This is not higher order, though it appears so.  I made this 
mistake myself many times.  That many rule systems do not support it is a 
seperate matter (and one we shall strive to address).

> The same problem arises for
> queries of the form <s,rdf:type,?c> (i.e., asking for all classes
> of a given instance). If you map triples of the form
> <s,rdf:type,c> to c(s), you again need higher order.

Again, you need to do something, yes.  But it is not a-priori higher 
order.

Since RIF is an interchange language, I'm pretty sure we are going to need 
a syntax and semantics like RDF that supports quantification over 
predicates.  Seems to me we want to enable the "mappings" between 
different languages to be expressed as declaratively as possible.  We 
shall see.

> Additional examples can be constructed with
> rdfs:subClassOf and rdfs:subPropertyOf.
> 
> The whole problem is somehow related to supporting T-box vs.
> A-box reasoning: for the straightforward mapping, Horn logic
> supports A-box reasoning, but not (much of) T-box reasoning.

I don't think so.  T-box vs. A-box reasoning is another issue, I'm sure, 
which overlaps a little with these kinds of problems.  This issue is 
accurately described in the RDF Compatibility page as "meta-modelling" - 
the ability to quantify over predicates and thus axiomatize them.

> > I think the syntax "triple(s,p,o)" would in this case be the more
> > feasible, since we decided to stay syntactically in function-free Horn
> > (with rule safety) for phase 1. 
> 
> Hmmm, the minutes of the Dec 20 telecon say:
> 
>    "digression clarified that phase 1 is full horn, not function-free
>     horn"

Right - I keep making this mistake myself.  Phase 1 is full horn.

-Chris
Received on Monday, 9 January 2006 20:44:42 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Tuesday, 2 June 2009 18:33:26 GMT