[Fwd: Re: Invitation for review of POWDER documents (Last Call)]

My intention is to answer 'yes' to all the points, ie, that the WG is
satisfied. Any objections to that?

Ivan

-------- Original Message --------
Subject: Re: Invitation for review of POWDER documents (Last Call)
Date: Mon, 27 Apr 2009 14:12:40 +0300
From: Stasinos Konstantopoulos <konstant@iit.demokritos.gr>
To: Ivan Herman <ivan@w3.org>
CC: Phil Archer <phil@philarcher.org>, W3C OWL Working Group
<public-owl-wg@w3.org>
References: <49D9D592.9030201@philarcher.org> <49E6E3DA.3080501@w3.org>

Ivan, W3C-WG, hi.

On Apr 16, 2009, at 10:52 AM, Ivan Herman wrote:

> - The reference should be to XSD1.1 and not XSD2:
>      http://www.w3.org/TR/xmlschema11-2/
>
> - 'At the time of writing, the OWL-2' should say "OWL 2" (ie, no  
> hyphen)
>
> - The reference to OWL 2 currently points to the OWL 2 Primer. We  
> think
> it would be better if it pointed at the (new) OWL 2 Document Overview:
> http://www.w3.org/TR/owl2-overview/

All updated, thank you.

> - The semantic condition refers to rdfs:Resource for the domain of
> hasIRI. Although the description refers to an extension of the RDF
> semantics, it makes use of, say, owl:DatatypeProperty. Hence it may be
> stylistically better to refer to owl:Thing.

I am leaning towards removing the domain triple altogether, as
it is obviously gratuitous.

> - The encoding of the condition in the example has several problems,
> partially due to some recent changes in OWL 2. These are:
>
>    - namespace changes (OWL 2 refers to xsd:pattern directly and not
> owl:pattern (OWL 2 reuses rdfs:Datatype instead of owl:datarange)

Updated.

>    - we also think that the type of restriction used is inappropriate.
> owl:hasValue should refer to a single individual and not to a
> datatype/datarange. Based on the rest of the POWDER semantics, what  
> you
> probably have to use is owl:allValuesFrom, but this is something you
> have to decide, of course

Shouldn't it be owl:someValuesFrom to guarantee that the specified
value exists? Since hasIRI is functional, it also guarantees that all
values are also as expected. I am interested in OWL WG's reaction to
this.

>    - the RDF mapping of facets is based on a list of blank nodes
> instead of the approach used in the current code
>
> The first example (the second has similar structure) should look
> something like:
>
> <owl:Restriction>
>  <owl:onProperty rdf:resource="....#hasIRI"/>
>  <owl:allValuesFrom>
>    <rdfs:Datatype>
>      <owl:onDatatype rdf:resource="...#string"/>
>      <owl:withRestrictions rdf:parseType="Collection">
>        <rdf:Description>
>          <xsd:pattern rdf:datatype="...#string">PATTERN</xsd:pattern>
>        </rdf:Description>
>      </owl:withRestrictions>
>    <rdfs:Datatype>
>  </owl:allValuesFrom>
> </owl:Restriction>

Indeed, modulo the owl:allValuesFrom vs. owl:someValuesFrom issue.

Best,
Stasinos

-- 

Ivan Herman, W3C Semantic Web Activity Lead
Home: http://www.w3.org/People/Ivan/
mobile: +31-641044153
PGP Key: http://www.ivan-herman.net/pgpkey.html
FOAF: http://www.ivan-herman.net/foaf.rdf