W3C home > Mailing lists > Public > www-rdf-interest@w3.org > June 2004

RE: XSLT for literal mapping/conversion

From: Kirkham, Pete (UK) <pete.kirkham@baesystems.com>
Date: Fri, 04 Jun 2004 16:45:29 +0100
To: www-rdf-interest@w3.org
Cc: Jeremy Carroll <jjc@hplb.hpl.hp.com>, Damian Steer <damste@hplb.hpl.hp.com>
Message-id: <820DBA1A8ECA1D45A557AFD03CF4DEE06E4FED@glkms0015>


> Jeremy Carroll
> It's true that the transform must be additive rather than replacing
> triples; and the total of all the triples must respect the
> semantics of all ontologies.

That depends whether you want to adopt a monolithic model that applications take views on, or a many linked models approach. In many other systems where the existence of a property in one view of the model (the r2:date in applications that understand that ontology) implies the existence of a property in another view (the r1:date in applications that understand that ontology), the views are separate and the transformations constitute a different metalinguistic layer.

> It is not clear how many ontologies and schema are
> designed with this in mind. e.g. if ex:date is an
> owl:FunctionalProperty
> in the above then it seems as though we might be hosed, unless we
> introduce a new property eg:sameDateAs which does the right thing.

Yes, that is one of the reasons (others include 'bow-tie' inheritance) why the many-models-with-mappings approach was taken in UML (QVT), EXPRESS (EXPRESS-X), CWM (no transform language, application specific mappings to common UML model), SGML (Architectural forms), IGES (application specific mappings to common ER model), CORBA (application specific mappings to common interface language, including object identity and properties), ASN.1 (ECN language to map non-standard encodings onto fragments of standard encodings, which is probably closest to this case as it involves different encodings of the same abstract data value), and even the normal use of XSLT for mapping XML instances of one schema to another. That's a lot of successful interoperation and metadata applications. Not that there aren't possible work-arounds to fit transformations into a monolithic model space, but having everything available as RDF (possibly a big advantage) doesn't have to imply that all RDF assertions exist and are true in the same model space. Lumping both source, target and metalanguage together in the same RDF world could well cause much complication without adding any real benefit.


Pete

********************************************************************
This email and any attachments are confidential to the intended
recipient and may also be privileged. If you are not the intended
recipient please delete it from your system and notify the sender.
You should not copy it or use it for any purpose nor disclose or
distribute its contents to any other person.
********************************************************************
Received on Friday, 4 June 2004 11:46:03 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Monday, 7 December 2009 10:52:07 GMT