- From: Dave Beckett <dave.beckett@bristol.ac.uk>
- Date: Wed, 13 Nov 2002 11:09:37 +0000
- To: Patrick Stickler <patrick.stickler@nokia.com>
- cc: Jonathan Borden <jonathan@openhealth.org>, WWW-Tag <www-tag@w3.org>
>>>Patrick Stickler said: > I've tried repeatedly to get this functionality into RDF datatyping, > but for some reason or another, the WG does not seem persuaded that it > is necessary at the moment. > > You present a very compelling use case for this, and fortunately, the > pieces are all in place to support typed XML literals. It's mostly > a matter of allowing them. They're not fundamentally different from > literals of type rdfs:XMLLiteral. I do find this a bit more compelling than the use-case presented by Jonathan based on webont's postponed requirement for complex XSD datatypes (If I recall it correctly). The difference is going from a one special datatype for XML content to a whole universe of them, thus requiring a new isXML flag, so that in RDF syntaxes written in XML, the XML content can be rendered back into elements and attributes. This doesn't happen unless you know it is XML. If any datatype URI-reference could potentially be XML and required to be written as such, you will need the isXML flag added to all typed literals. > I suggest you ask the RDF Core WG to open this issue. I ask you don't since it requires an RDF model change (adding an isXML flag to typed literals) as well as changing RDF/XML again. You underestimate the impact of such "simple" changes. Dave
Received on Wednesday, 13 November 2002 06:11:13 UTC