Re: proposal: Structured Datatypes

Jim Hendler wrote:


> Jonathan,
>   In our work on dynamic composition of web services using semantics,
> we've really had to face this issue again and again -- we need to go
> from semantic constructs (typically an instance of a class and its
> associated properties) to an XML data structure -- I've become
> convinced this is another one of those problems thta looks easy if
> you're mainly looking at one aspect of it, but when you get into
> details, it looks like a very general purpose mechanism, more complex
> than what you propose, will be both needed and desired.

Yes, what I propose is a simple first step that is not intended to be a
complete solution but I think it might allow enough of a jumpstart to get a
good idea about exactly what a complete solution would (sic) entail.

 I also worry
> that this is really something that RDF Core should do (or not do) and
> for us to have a owl-specific solution makes me nervous w/respect to
> the interaction of OWL and RDF (particularly, I think these complex
> types will be instantiated in instances, but the instances typically
> would use the RDF as opposed to OWL MT in general)

Yes well according to our *requirements*
http://www.w3.org/TR/webont-req/#section-requirements

[[
R15. Complex data types
The language must support the definition and use of complex/ structured data
types. These may be used to specify dates, coordinate pairs, addresses, etc.
]]

which is motivated by our multimedia use case
http://www.w3.org/TR/webont-req/#usecase-multimedia

it seems to me that structured datatypes fall squarely within the charter of
WebOnt. I mean, gasp, if this is something you are not comfortable doing,
why would this be an OWL *requirement*

besides, RDFCore *isn't* dealing with XML dataypes (the ones that have XML
in them a.k.a. XML Schema complexType) and as you say this is an issue that
we will continue to face over and over until it is properly dealt with. An
ideal solution would model an XML Schema particle as an owl:Class, but at
the very least I'd like to be able to talk about pieces of XML e.g. say that
this is the sameIndividualAs that or that this XML type is the sameClassAs
that etc.

Jonathan

Received on Wednesday, 6 November 2002 22:58:30 UTC