Namespace dispatching

I think I understand from the IRC logs that the idea is that the world
of XML documents will be divided into two sets, those where the
top-level namespace allows dispatching and those where it does not. The
latter documents must not be shipped with a +xml suffix. 

I am uncomfortable with the idea that these XML documents will not be
self-describing. Whereas one can almost always detect the Unicode
encoding, and can almost always detect the top-level namespace, one
cannot reliably know how to dispatch without metadata. Top-down
self-descriptiveness is one of the major advantages of XML and I think
that doing otherwise should be deprecated.

In my experience, this ability to have a top-level namespace be other
than the controlling one is at best a syntactic shortcut. No expressive
ability would be lost by inventing a wrapper element that did nothing
more than say: "I'm using XSLT and the thing inside me is a literal
result element". i.e.

<xslt:literal>
 <html>
 </html>
</xslt:literal>

I propose that the namespaces spec be updated to respect this new view
of the "meaning" of the top-level namespace and that future versions of
W3C specs like XSLT be aligned to conform. If it helps, we could call
this a new issue: 

"Which of the following are appropriate triggers for determining the
document type of an XML document when metadata is unavailable:

 1 DOCTYPE statement
 2 top-level namespace
 3 schema reference declaration
 4 other root-level declared namespaces
 5 any attribute on the root element
 6 anything in the document"

I have not seen a compelling use-case for anything other than 1 and 2.
Most of the others are just typing conveniences, IMO.

 Paul Prescod

Received on Tuesday, 19 February 2002 09:05:48 UTC