[XML-DEV Mailing List Archive Home] [By Thread] [By Date] [Recent Entries] [Reply To This Message]

RE: OWL for the XML layer

• To: "Lyndon J B Nixon" <nixon@f...>,"xml-dev" <xml-dev@l...>
• Subject: RE: OWL for the XML layer
• From: "Danny Ayers" <danny666@v...>
• Date: Fri, 22 Nov 2002 19:33:43 +0100
• Importance: Normal
• In-reply-to: <3DDE2553.7090308@f...>

>>>NB One sentence summary: I can assert myrdf:ZipCode owl:samePropertyAs
>>>yourrdf:PostCode
>>>but why can't I also assert  myxml:ZipCode owl:sameDatatypeAs
>>>yourxml:PostCode.
>>>
>>>
>>
>>You can assert it, but what does it mean?
>>
>that the element ZipCode in my XML Schema has the same semantic meaning
>as the element PostCode in your XML Schema.
>If I understand correctly, OWL only allows this to be done if ZipCode
>and PostCode are RDF properties. However
>the differentiation that XML/XMLS is purely syntactical and RDF/RDFS
>semantic is false; XML vocabulary authors
>associate human-understandable meaning to their schemas (most humans
>will understand what is meant when an address contains a XML element
>called ZipCode or PostCode). But how to extend this semantic meaning to be
>machine processable, if the Web ontology languages won't permit it?

Ok, my original response was rather flippant and I apologise. But the
semantics of myxml:ZipCode are (presumably) given by the XML schema and
whatever else kind-of-intentional semantics you have given it in human
terms. Ok, so a key part to the utility of myrdf:ZipCode comes from the
human part, but OWL does understand the RDF language in which the
machine-explicit semantics are expressed. This isn't the case for
myxml:ZipCode. There are certainly potential (probably implemented) mappings
between the XMLS semantics and the RDF semantics, and there has been work
done on using XMLS datatyping within RDF. There is also potential for the
direct interpretation of XMLS-backed terms in OWL along the lines you are
suggesting. This would be nice to see, and could be extremely useful but I
suspect it could be very hard work. There are likely to be quite a few
obstacles to overcome before your statement could be used in practice. I
think a lot of these obstacles would stem from the syntactical similarities
of the two systems - a (very loose) analogy might be mixing Javascript and
Java. The first obstacle that come to mind is - how does the interpreter of
the (RDF) document know that XMLS semantics apply to terms in the myxml
namespace and RDF semantics apply to myrdf terms?

There is at least one possible simple way forward on that - for the author
to state *outside* of the XSD/RDF schemas that myxml:ZipCode
owl:sameDatatypeAs yourxml:PostCode, and then from then on the term has the
required meaning within any OWL statements.

Cheers,
Danny.

• References:
  • Re: OWL for the XML layer
    • From: Lyndon J B Nixon <nixon@f...>

• Prev by Date: Re: The non-future of XHTML x.x [LONG]
• Next by Date: Re: Re: Stalled browser development...
• Previous by thread: Re: OWL for the XML layer
• Next by thread: Re: OWL for the XML layer
• Index(es):
  • Date
  • Thread