- From: Jonathan Borden <jonathan@openhealth.org>
- Date: Wed, 12 Jun 2002 18:03:53 -0400
- To: "Joshua Allen" <joshuaa@microsoft.com>, "Tim Bray" <tbray@textuality.com>, <www-tag@w3.org>
Joshua Allen wrote: > > > > 1. Type-augmented XML is a good thing and a recommendation should be > > prepared describing it both at the infoset and syntax level. (I gather > > Type-augmented XML needs a type system. Which are you recommending? > A) Single spec, which uses one existing type system (XSD, WebOnt, XDR, > RDFS, etc.) > B) One spec for each > C) Single spec that combines many existing type systems > D) Yet another type system If we consider an XML type as equivalent to a class, or as generally defining constraints such that the set of XML fragments which conform to the constraints are the instances of the set, then XML can be considered to have a single "type system" with different ways to specify constraints. For example, the XQuery formal semantics type language can _easily_ incorporate both XML Schema types and RELAXNG "patterns" -- witness the, err, similar syntax with RELAXNG non-XML. Integrating these XML type systems with RDF and WebOnt/OWL is on the WebOnt issues list. see: http://www.openhealth.org/WOWG/IssueStructuredDatatypes.html and RDF: http://www.coginst.uwf.edu/~phayes/RDF_Datatyping060102_draft.html#ntoc_2.1 Peter Patel-Schneider and Jerome Simeon's perspective that incorporates XML Schema tightly into the Semantic Web vision: http://www-db.research.bell-labs.com/user/pfps/talks/building-on-xml/ So, at the end of the day, we don't necessarily need an entirely new type system, just a way to identify types. The wrinkle that needs to be sorted out is whether we ought identify types by: a) QNames, as in XML Schema b) URI refs as in RDF(S)/OWL. and how to meld these mechanisms -- well I've already raised this issue, and perhaps by reading the above references you will see why I, for one, and hopefully not the only one, think this is an important issue to resolve. > 4. Work on XQuery and other things that require a Type-Augmented > Infoset > > must not depend on schema processing, and should not have normative > > linkages to any schema language specifications. > > Are you saying that the XQuery type-augmented infoset should be in a > separate spec (which might have normative linkage to XSD), or that the > XQuery should be changed to allow *any* type-augmented infoset spec that > someone chooses to implement? > At the very least, a type could be specified by a QName or its URI ref equivalent. Jonathan
Received on Wednesday, 12 June 2002 18:10:10 UTC