- From: Thomas B. Passin <tpassin@home.com>
- Date: Tue, 2 Oct 2001 17:39:19 -0400
- To: <www-rdf-logic@w3.org>
[Peter F. Patel-Schneider] > From: "Thomas B. Passin" <tpassin@home.com> > Subject: Re: Literals (Re: model theory for RDF/S) > Date: Tue, 2 Oct 2001 16:37:17 -0400 > > > [Peter F. Patel-Schneider] > > [...] > > > For example, the syntactic equivalence of two URIs can be determined > > according to the normalization rules for URIs, and clearly belongs in RDF. > > RDF is supposed to have a datatype system, so it's reasonable to expect it > > to be able to tell strings from integers, but you can't apply that to > > literals because they can't be the subject of statements (unless that rule > > gets changed). > > Also agreed, but I sense that there are quite a number of people that are > not in favour of RDF having a datatype system. > Yes, and let me walk through in the same way as before if RDF were to end up with no type system. With no datatype system, RDF can make statements about anything but literals, but not any and all statements (no statements about datatypes) unless you go outside RDF proper to another layer. With no ontology, there is no basis for any inferencing activity, and RDF becomes essentially a syntax for producing (take your choice) edge-labeled graphs or sets of triples. This would be simpler and cleaner, with RDF being purely structural. Multiple type and inferencing systems could be plugged in on top. But you would have to have such a layer. Does this capture it? Cheers, Tom P
Received on Friday, 5 October 2001 18:21:58 UTC