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

Re: question about "xsltTransformation" property in DAML-S 0.9+

From: David Martin <martin@ai.sri.com>
Date: Tue, 10 Jun 2003 22:53:53 -0700
Message-ID: <3EE6C3F1.41AC4432@ai.sri.com>
To: Joachim.Peer@unisg.ch
CC: www-ws@w3.org

Joachim -

Thanks (belatedly) for these comments.  I realize Massimo has already responded
to selected points, but I'd like to add one brief comment ...

Joachim.Peer@unisg.ch wrote:

> dear DAML-S researchers,
>
> i would like to compile a brief comparison of the various existing
> approaches for mapping between DAML+OIL(OWL) and XML, and of course i am
> very interested to learn about the mapping concept envisioned for DAML-S
> 0.9
>
> Does there exist an example/explanation of the "xsltTransformation"
> property, or is this work in progress?
> To discuss the various problems (and solutions) of the XSLT based mapping,
> I have written a (very simple) example,  a mapping of a DAML+OIL  "purchase
> order" to some XML represention.
>
> The example is available online:
> http://sws.mcm.unisg.ch/mapping/po.daml (the DAML+OIL ontology; the XML
> schema is http://www.w3.org/TR/xmlschema-0/#po.xsd)
> http://sws.mcm.unisg.ch/mapping/po.xml  (an instance of the XML schema)
> http://sws.mcm.unisg.ch/mapping/po_instance.rdf  (an instance of the
> DAML+OIL ontology)
> http://sws.mcm.unisg.ch/mapping/PO-daml2xml.xsl  (the XSL stylesheet which
> maps from ONE RDF/XML representation to XML)
>
> Of course i do not know if this is compatible with the current ideas
> regarding the "xsltTransformation" property ;-)
>
> However, the example illustrates some open issues/some of the questions i
> currently have:
>
> *  a DAML+OIL concept can have multiple RDF/XML representations => Q: Do
> you plan to impose some guidelines which restrict the use of certain RFD
> abbrev. syntaxes, or do you think it's the responsibility of the softare
> agents, to figure out how _exactly_ the mapping should be carried out ,
> e.g. to figure out which RDF/XML style is used,  etc.

It's a good point.  We've only talked about this a little in the DAML-S
Coalition.  The consensus seems to be that it's not practical to expect an XSLT
transformation to handle *any* possible XML representation of the relevant
DAML+OIL concepts - so it probably would be useful to impose some guidelines.
However, no recommendation is currently being made about this.  One possible
set of guidelines would be the OWL XML presentation syntax:


http://lists.w3.org/Archives/Public/www-webont-wg/2002Dec/att-0295/01-OWL-XML-Schemas.html

but I'm not entirely clear yet whether this is sufficiently restrictive to
address the problem.

Regards,
David Martin


>
> * XSLT Transformations are not bi-directional => Q: will the stylesheets
> for inputProperties look different than those for outputProperties?
>
> thanks in advance,
> joachim
>
> Joachim Peer
> Research Assistant
> MCM Institute, University of St. Gallen
> Blumenbergplatz 9, 9000 St. Gallen, Switzerland
> Phone: ++41 (0) 71 224 3441, Fax: ++41 (0) 71 224 2771
Received on Wednesday, 11 June 2003 01:54:03 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Tuesday, 3 July 2007 12:25:43 GMT