Michael Champion wrote:

> Nobody I know is arguing anything remotely resembling "everybody must
> use the same data model to process XML".  The point of cleaning up the
> data model mess would be to rectify the ugly mismatches between DOM
> and XInclude, DOM and XPath, etc., i.e. make simple things much
> simpler for novices, not make life harder for the  uber-geeks who are
> comfortable with the current situation.

That's been tried. It's called the XML Infoset, and all it did was add 
one more slightly incompatible data model to the mix. Nobody, including 
uber-geeks, is comfortable with the current situation, but I'm afraid 
it's too late for this one. The horse has left the barn, and the barn 
burned down. We live in a world of many data models that are similar but 
not quite the same around the edges. Annoying sometimes, but not fatal.

-- 
Elliotte Rusty Harold  elharo@m...
XML in a Nutshell 3rd Edition Just Published!
http://www.cafeconleche.org/books/xian3/
http://www.amazon.com/exec/obidos/ISBN=0596007647/cafeaulaitA/ref=nosim