OASIS Mailing List ArchivesView the OASIS mailing list archive below
or browse/search using MarkMail.


Help: OASIS Mailing Lists Help | MarkMail Help



   RE: [xml-dev] Ten new XQuery, XSLT 2.0 and XPath 2.0 Working

[ Lists Home | Date Index | Thread Index ]
  • To: "Dave Pawson" <dpawson@nildram.co.uk>,<xml-dev@lists.xml.org>
  • Subject: RE: [xml-dev] Ten new XQuery, XSLT 2.0 and XPath 2.0 Working
  • From: "Michael Rys" <mrys@microsoft.com>
  • Date: Fri, 9 May 2003 11:20:35 -0700
  • Thread-index: AcMWVBv/hSleBVA9TDSAxXDtKeffsgAAZgEw
  • Thread-topic: [xml-dev] Ten new XQuery, XSLT 2.0 and XPath 2.0 Working

If you are satisfied with just "untyped" XML, then XPath/XQuery should
work fine. You get a weakly typed language that performs implicit
casting to the types you use and has default type promotion rules when
its built-in operators have only untyped values as its arguments.

Instead of complaining about the design at the "I don't like XSD, thus I
don't like XPath 2.0/XQuery 1.0"-level, I would recommend that you give
us feedback on the implementation of this principle. I am aware of one
bug so far that we work on related to the static typing option and
xs:anyType. More feedback on what you want to do but cannot do would be

One of my personal design goals is to have XPath/XQuery strongly typed
for typed instances and weakly typed for untyped instances and all work
inside the same language framework. I think we have done a pretty good
job at that given the constraints imposed on us by XSD and other design
goals [1]. 

And note that this makes the specification more complex, but the
language actually simpler to use if you constrain yourself to
infoset-only data.

Best regards

[1] Yeah, I know: I am biased and can't see beyond the Working group
anymore. Etc pp...

> -----Original Message-----
> From: Dave Pawson [mailto:dpawson@nildram.co.uk]
> Sent: Friday, May 09, 2003 10:55 AM
> To: Michael Rys; xml-dev@lists.xml.org
> Subject: RE: [xml-dev] Ten new XQuery, XSLT 2.0 and XPath 2.0 Working
> At 10:44 09/05/2003 -0700, Michael Rys wrote:
> >The data model also can be generated from just Infosets. The
> >will be annotated with xdt:untypedAtomic for attribute values and
> >(currently) xs:anyType (soon xdt:untypedAny or similar) for element
> >content type.
> >
> >So if XSD goes down the drain (what does this mean anyway?)
> Shrivels and dies through lack of use, or is overtaken by
> a schema that works.
> >  XSLT and
> >XQuery still work on untyped data.
> I remain to be convinced. I've only had a few hours with Saxon 7.4.
> XSLT is my concern today.
>    I've no idea what xquery want..
> >And there is no reason why you cannot add your own built-in type
> I really don't want to Michael.
>    The xslt+xpath1 'data model' is adequate for me.
> regards DaveP


News | XML in Industry | Calendar | XML Registry
Marketplace | Resources | MyXML.org | Sponsors | Privacy Statement

Copyright 2001 XML.org. This site is hosted by OASIS