Re: Enlightenment via avoiding the T-word

[Rick Jelliffe <ricko@allette.com.au>]

 From: "Francis Norton" <francis@redrice.com>
 
> Rick Jelliffe wrote:
> 
> > or, more likely, you have
> >   <xsl:template match="x:name">
> >      <xsl:if select="parent::x:person">
> > 
> It's not even that simple. x:person may be a local element. How far back
> up the tree do you need to go to get to an element whose name maps
> uniquely to a single content model? All the way, potentially. That's why
> I'd be interested in doing a meta-transform that reads in a schema and
> generates, not a schema validator, but an element-labeller that would
> mark up every element in a valid message with a unique type attribute.

This is the architectural forms route. xsi:type.

Another approach would be a tool that checks through scripts and, 
when an element has a local type, flags or revises the code to ignore
the new type. Yuck..

> I've got one that will do some of this - but it can't handle redefines
> or anonymous complexTypes. I think I'm going to have to read the formal
> description for a handle on naming anonymous types. (It also goes
> blazingly fast because it uses modes - transforming a 2K message against
> a 600K labelling transform takes 1 ms in MSXSL on my laptop)

Does this include the timing for downloading the new schema, and transforming
it into your type annotator?  

> > Push programming and pull programming are both common methods with XML.
> > 
> Yes, and given local types, push transforms can only be supported by
> having access to at least some subset of the PSVI. I think the approach
> above might hit the 20/80 point.
 
Which is to say that local types complicate push-processing and require unavailable
and uninvented technology. 

Cheers
Rick Jelliffe