- From: Gerd Wagner <wagnerg@tu-cottbus.de>
- Date: Mon, 16 Oct 2006 20:40:32 +0200
- To: "'Dan Connolly'" <connolly@w3.org>
- Cc: "'Chris Welty'" <cawelty@frontiernet.net>, "'Public-Rif-Wg \(E-mail\)'" <public-rif-wg@w3.org>
> > As an example of typing in a rule condition language,
> > consider the following property atom in IRL:
> >
> > cust:GoldCustomer; sCart:ShoppingCart(customer == cust);
> >
> > It states that the customer property of the shopping cart sCart
> > has the value of cust, which is a varaible of type GoldCustomer.
> > It involves two variables, cust and sCart, both of which are
> > typed. Having a specific form, it should also be viewed as a
> > special type of atom, namely an object-valued property atom.
> > Rewritten in the current core condition language proposal as
> > the untyped standard predicate logic atom
> >
> > customer( sCart, cust),
>
> I would expect that sort of type stuff* to turn into additional
> premises.
>
> In SPARQL[1], I might write
>
> ... WHERE { ?cust a ex:GoldCustomer.
> ?sCart a ex:ShoppingCart;
> ex:customer ?cust }
Sure, you can map it in this way, but I don't see how you can
reconstruct the typing information at the interchange destination
side easily from such a reduced representation. How can you
reconstruct, e.g., the destination SWRL atom
customer( I-variable(sCart) I-variable(cust))
from the following concjunction of three atoms
ShoppingCart( sCart) & Customer( cust) & customer( sCart, cust)
in a language where monadic predicates do not necessarily express
types?
-Gerd
Received on Monday, 16 October 2006 18:40:56 UTC