context (was Re: ANN: SAX Filters for Namespace Processing)

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

On 01 Aug 2001 11:27:12 +0100, Sean McGrath wrote:
> I remember the early days of namespace rec. development when people
> where talking about namespaces having no effect on parsers, no effect
> on apis etc. I find it very depressing to look at the
> increase in complexity of fundamental subsystems such as
> SAX/DOM/XPATH/Query etc.etc. because of namespaces.

I agree completely, as I agree with most of Sean's message.

> Context is always a difficult problem. Namespaces writes
> the context problem in stone and makes it an
> "in your face" aspect of pretty much everything you
> might want to do with XML documents.

One thing that's interesting to me here is the contrast between the use
of context in these XML Namespace/W3C XML Schema models and RELAX/RELAX
NG. 

RELAX and RELAX NG provides some very cool context-sensitive document
modeling tools which let you say things like "paragraphs inside of lists
have a different content model than paragraphs in the main body of the
document."  (I believe that TREX, Schematron, and Examplotron are
similarly capable, and have no problems with things like substitution
groups muddying context.)

That understanding at validation doesn't make any claims about the scope
of meaning in the document, however.  You don't need to do RELAX
processing to figure out what the components in the document mean.  It's
a far cleaner separation of validation processing from
semantics-by-implication.

I have to admit that I'm hoping the current situation with regard to XML
Namespaces and W3C XML Schemas is a giant practical joke, but I see no
signs of pranksters coming forward with a gleeful smile to announce that
they were just kidding.