W3C home > Mailing lists > Public > www-webont-wg@w3.org > June 2003

Re: OWL DL in RDF

From: Sean Bechhofer <seanb@cs.man.ac.uk>
Date: Wed, 11 Jun 2003 12:30:23 +0100 (GMT Daylight Time)
To: Jeremy Carroll <jjc@hpl.hp.com>
cc: www-webont-wg@w3.org
Message-ID: <Pine.WNT.4.44.0306111201060.1840-100000@potato>

On Wed, 11 Jun 2003, Jeremy Carroll wrote:

>
>
> Review of
> http://wonderweb.man.ac.uk/owl/rdf.shtml
> dated
> 05/30/03
>
> Scope of review: purely technical - I believe the WG is already convinced of
> the stylistic merits of this work.

Thanks Jeremy.

> Footnote [3]: suggest add:
> --
> x rdf:type rdfs:Class
> x rdf:type owl:Class
> x rdf:type owl:Restricition
> --
> as one of the possibilities

Done

> Restrictions
> ==========
>
> Suggest s/involving/with predicate/
> (leaving 'involved' to indicate the object of a triple)

Done

> "It is not the subject of any other triples"
> Unfortunately false, and hardish to fix.
> 1) see footnote 3 fix
> 2) it may be the subject of many owl:equivalentClass or owl:disjointWith
> triples.

I've tried to fix this with the following:

[[
o It is not the subject of any other triple with predicate rdf:type (ref
to footnote 3).

...

o Any other triples in which x is the subject should have predicate
owl:equivalentClass or owl:disjointWith.
]]

> Class Axioms
> ===========
> weak suggest s/|/\nor\n/

Done.

>
> Property Axioms
> =============
> owl:DatatypeProperty can have an rdfs:domain

Done.

> rdfs:range suggest
> s/specifying a data range/specifying a data range with type owl:DataRange/
> also note rdfs:Literal can be specified as a range

Done

> Boolean Class Expressions
> ======================
>
> suggest adding at end,
> if the owl:Class is unnamed then only one such expression is permitted.
> Named classes can have any number of these expressions.

Added:

[[If the owl:Class is a bnode (i.e. the class is unnamed), then it can
only be the subject of at most one triple with predicate
owl:intersectionOf, owl:unionOf or owl:complementOf. If the class is
named, any number of such triples are allowed. ]]


> Avoid Structure Sharing
> ===================
>
> s/AS&S/S&AS/

Done (and earlier in the document too).

> Avoid Orphan bnodes
> ==================
>
> unnamed individuals are forgotten
>
> suggest
>
> "In general, bnodes occurring in the graph either represent unnamed
> individuals or should "

Done


> Omission
> ========
>
> owl:equivalentClass and owl:disjointClass suggest waiting to see the B1 B2
> discussion before drafting

Good idea!

An updated version (with additional minor formatting changes to use the W3
styles) is now at:

http://wonderweb.man.ac.uk/owl/rdf-03-06-11.html

Cheers,

	Sean

-- 
Sean Bechhofer
seanb@cs.man.ac.uk
http://www.cs.man.ac.uk/~seanb
Received on Wednesday, 11 June 2003 07:32:15 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Monday, 7 December 2009 10:58:01 GMT