W3C home > Mailing lists > Public > w3c-rdfcore-wg@w3.org > April 2002

Schema. (fwd)

From: Jan Grant <Jan.Grant@bristol.ac.uk>
Date: Fri, 26 Apr 2002 11:23:47 +0100 (BST)
To: RDFCore Working Group <w3c-rdfcore-wg@w3.org>
Message-ID: <Pine.GSO.4.44.0204261120260.23534-100000@mail.ilrt.bris.ac.uk>
Dan asked me to send this on-list. Mostly niggles.

	The RDF Schema class and property system is similar to the type systems
	of object-oriented programming languages such as Java.

I think this is disingenuous. It's not very like it at all,
_particularly_ not like Java. Practices in schema and ontology creation
(eg, making new superclasses when old subclasses are already
inexistence; general frivolous attitude towards cycles) aren't common in
most OO languages.

	rdfs:Resource 		The class Resource.

Say more?

	 rdfs:Class 		 The concept of Class
	 rdf:Property 		 The concept of a property.
	 rdfs:Literal		 This represents the set of atomic values,
				 eg. textual strings.
	 rdf:Statement 		 The class of RDF statements.
	 rdfs:Container 	 This represents the set Containers.
	 rdf:Bag 		 An unordered collection.
	 rdf:Seq 		 An ordered collection.
	 rdf:Alt 		 A collection of alternatives.
	 rdfs:ContainerMembershipProperty
				 The container membership properties,
				 rdf_1, ... rdf_n, all of which are
				 sub-properties of 'member'.

We've got concepts, classes and sets here. I know it's really tedious to
repeat the same phrase over and over, but people might ask: "do RDF
statements form a class? why not a set? What's this set Containers? etc."

rdfs:Class	The class of all _rdfs:Class_es
rdf:property	etc.

Literal: don't say "atomic". Atomic means "indivisible"; literals (even
the simple ones we're working with) have structure. That structure
might not be expanded in terms of nodes and arcs, but it's wrong to say
"the structure doesn't appear in the graph". It _does_ appear, but it's
bundled up into a literal node. Call it "The Class of literal values"
and point to the MT for a definition.

The only other thing is that I'd avoid rdf:_n unless you can italicise
or otherwise show that "rdf:_n" isn't one of these.


I'd also (I know it's late in the game) be inclined to float the
suggestion of "fixing" the rdfs:subClassOf & co. properties, deprecating
these in favour of names that are the right way around according to M+S,
and including entailment rules to show that they're the same thing.
(just make them circularly subproperties of their renamed versions)
- particularly if we're switching namespace.


-- 
jan grant, ILRT, University of Bristol. http://www.ilrt.bris.ac.uk/
Tel +44(0)117 9287088 Fax +44 (0)117 9287112 RFC822 jan.grant@bris.ac.uk
Whose kung-fu is the best?
Received on Friday, 26 April 2002 06:24:08 EDT

This archive was generated by hypermail pre-2.1.9 : Wednesday, 3 September 2003 09:47:37 EDT