- From: Peter F. Patel-Schneider <pfps@research.bell-labs.com>
- Date: Wed, 16 Jan 2008 13:40:38 -0500 (EST)
- To: jjc@hpl.hp.com
- Cc: public-owl-wg@w3.org
From: Jeremy Carroll <jjc@hpl.hp.com>
Subject: ISSUE-29 datarange and plain literals with language tags
Date: Wed, 16 Jan 2008 17:56:08 +0000
> I have just had a thought.
>
> Owl:DataRange is also used for sets of plain literals with language
> tags, as well as for sets of typed literals.
>
> I am unclear whether rdfs:Datatype can be used for sets of plain
> literals with language tags.
>
> Jeremy
Should still work, as instances of rdfs:Dataype don't have to be RDF
datatypes.
The semantic conditions and axiomatic triples wrt rdfs:Datatype are:
1/ If x is in ICEXT(I(rdfs:Datatype))
then <x, I(rdfs:Literal)> is in IEXT(I(rdfs:subClassOf))
2/ rdf:XMLLiteral rdf:type rdfs:Datatype .
rdfs:Datatype rdfs:subClassOf rdfs:Class .
3/ If <aaa,x> is in D (i.e., D is an RDF datatype)
then I(aaa) is in ICEXT(I(rdfs:Datatype))
The axiomatic triples in 2/ are unproblematic.
The semantic rule 3/ talks about RDF datatypes, and thus doesn't matter,
as the dataranges are not RDF datatypes.
The semantic rule 1/ says that instances of rdfs:Datatypes are
subclasses of rdfs:Literal so we do have to be careful. However,
rdfs:Literal contains ... wait for it ... all RDF literals, which,
includes ... wait for it again ... plain literals.
So, saying that { "1" "3"@en } is an instance of rdfs:Datatype is
unproblematic.
peter
Received on Wednesday, 16 January 2008 19:11:21 UTC