Lists Home |
Date Index |
At 04:17 PM 12/4/2002 -0500, Norman Walsh wrote:
>-----BEGIN PGP SIGNED MESSAGE-----
>/ John Cowan <firstname.lastname@example.org> was heard to say:
>| Again, the mere *provision* of typing metadata does not prevent such reuse.
>| However, if *standard tools* assume that the metadata is sound, then
>| transgressive reuse may indeed be made difficult. Obviously, purely lexical
>| tools are not affected, but tools based on XQuery/XPath2/XSLT2 will not
>| be purely lexical in this sense (whereas XPath1/XSLT1 are).
>Taking the specific case of XQuery/XPath2/XSLT2, I'm not sure I see
>the problem. Given
><foo n="1">Network Drive</foo>
>I might write a template that matches those elements in a purely
> <xsl:template match="baz">...</xsl:template>
I agree with Norm.
XQuery is a strongly typed language, but it is designed to allow untyped
and typed data to be used together gracefully. In XQuery, you do have to
cast if data has a datatype that prohibits a specific kind of operation,
but generally not if the data is untyped. The type information available in
a query depends greatly on whether schemas have been imported into the query.
XQuery can work on well-formed XML, XML governed by DTDs, or XML governed
by XML Schema. One query can involve data from all three kinds of XML.
I think John Cowan's text implies that there might be problems with reuse
if there are datatypes in the data and XQuery is being used to process it.
I would be interested in seeing some concrete examples that illustrate this