RE: ANN: SAX Filters for Namespace Processing

["Simon St.Laurent" <simonstl@simonstl.com>]

On 31 Jul 2001 14:11:37 -0700, Evan Lenz wrote:
> > 1 rule. For an XML document to be namespace compliant, it will not contain
> > any unqualified elements once any namespace is in scope.
>
> ...which would completely break XSLT literal result elements which represent
> a pretty intuitive use case for unqualified child elements.

XSLT is an interesting case.  While the unqualified elements in some
sense "belong" to the surrounding qualified XSLT markup, they clearly
aren't treated as properties of the XSLT markup.  I'd strongly not
recommend using my filters on XSLT templates, and I'll add some
documentation to that effect.  (Exclusions make that easy to achieve
even in mixed contexts.)
 
> Also of note is the fact that XML Schemas was not the first technology to
> use QNames in attribute values. Whether good or bad, that approach was
> adopted by XSLT.

Was that in the original XSLT plans?  Or did that emerge as namespaces
developed and had to be encapsulated in XPath?  I can't remember.

XPath in XPointer with xmlns - now there's a demonstration of how much
fun QNames can be.