W3C home > Mailing lists > Public > www-tag@w3.org > March 2002

Re: Namespace dispatching

From: Paul Prescod <paul@prescod.net>
Date: Wed, 06 Mar 2002 17:04:38 -0800
Message-ID: <3C86BCA6.753F958C@prescod.net>
To: TAG <www-tag@w3.org>
Norman Walsh wrote:
> 
>...
> 
> Display all the MathML equations in this document. Display all the SVG
> diagrams. Present a concordance of Japanese words...

> | Most of the others are just typing conveniences, IMO.
> 
> This thread goes on to talk about nested wrappers and things as if all
> typing conveniences were just shortcuts for nested wrappers, but I
> don't think that's true. As for processing models, well, I think something
> dependency based is probably a better choice than nested wrappers.
> But I would say that, wouldn't I?[1] :-)

There are two different problems to be solved. 

1. a document receiver specifying a pipeline of instructions to generate
a multi-phase transformation of the document.

2. An author of a document indicating the ordering of transformations to
be applied in order to get at the real data in the document as opposed
to the markup tricks used to manage its separate bits and its embedded
namespaces.

The second is more urgent to me than the former because I can do the
former with Python or Perl or whatever. But I can't ship Python or Perl
code to Web clients and tell them: "execute this to get the order of
XSLT, XInclude and XBase processing correct." Your use cases like
"Display all the MathML equations in this document" indicate that XPipe
is about dealing with the former, but the namespace dispatching thread
was about the latter.

 Paul Prescod
Received on Wednesday, 6 March 2002 20:07:49 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Thursday, 26 April 2012 12:47:05 GMT