Re: Namespace: what's the correct usage?

["W. E. Perry" <wperry@fiduciary.com>]

Martin Gudgin wrote:

> Yeah, having XPath be type based rather than name based will be fantastic.
> The MSXML4 parser[1] from Microsoft has some ( proprietary ) support for
> that today. I can't wait for that stuff to be standardised and avaibable in
> all processors.

My god! Next I suppose you'll be wanting a post-schema-validation infoset, or
some such. ;-) But, seriously, how can XPath not be name- (or text- or syntax-)
based and remain XML? Whatever the markup may say (and what it says is
definitive), type does not inhere in the syntax or the 'data' conveyed by that
syntax:  it inheres (or more correctly, is elaborated) in the processed outcome
of that syntax. To achieve identical outcomes requires congruent processing, and
that is the one thing markup-based information handling cannot assume. XML is
syntax precisely to provide the broadest common denominator for admittedly
differing processes to operate upon. Shift the assumption to a homogeneity of
processing and while you may have the basis for an API you do not have XML or
anything that might be reasonably understood as expressing the fundamental
nature of markup.

Respectfully,

Walter Perry