Re: process order (still...)

> The ordering that I was trying to stress was different and
> implementation dependent is the `chronological' order in which nodes
> are added to the result tree. That is when the actual templates
> get fired. This may happen in any order, or all at the same time.

Ah, true, but we write stylesheets as if the templates are going to be
instantiated serially. "Add foo to the result tree, add bar to the result
tree, identify this node-set, execute the instructions in the template
that best matches the first node in the set, then the second node, then
the third, etc." If we didn't have rules for ordering, we wouldn't ever be
able to predict what the result tree would look like.

So are we in agreement when I assert that 'the nodes are processed in
document order' really means something more like 'the result tree and/or
result tree fragments created by the templates for these nodes are created
such that the spatial order of the nodes *is as if* the templates had been
instantiated one at a time, for each node in document order' ?

Not that that's any less confusing :)

   - Mike
___________________________________________________________
Mike J. Brown, software engineer, Webb Interactive Services
XML/XSL stuff: http://www.skew.org/    http://www.webb.net/


 XSL-List info and archive:  http://www.mulberrytech.com/xsl/xsl-list