Re: Namespaces, schemas, Simon's filters.

[Francis Norton <francis@redrice.com>]

Richard Tobin wrote:
> 
> There are two points in dispute: whether local elements are good at
> all, and whether unqualified ones are (no doubt there is also a view
> that *only* unqualified ones are good, but I haven't noticed anyone
> arguing that).
> 
OK - the case deserves to be made, even if at the expense of any claim I
may make to intellectual consistency.

[1]	it means all namespaced elements are uniquely typed by their names,
which reduces the complexity of implementing processors using non-PSVI
aware tools (obviously I'm thinking of XSLT / XPath here and below)

[2]	it makes it easier to process local elements since each local
element's first name-spaced ancestor will be of a type that is uniquely
determined by its name

I think that the case starts to fall down if there is any expectation
that instances will be mixed in to documents with alien elements.

Francis.