Re: Namespaces, schemas, Simon's filters.

[Ronald Bourret <rpbourret@rpbourret.com>]

"Fuchs, Matthew" wrote:

> I think this is close, certainly closer than some are willing to
> acknowledge.  One important thing is, you don't distinguish between valid
> and well-formed.  You are basically looking at things from a validation, or
> well-typed, perspective - as do I.  Evan and Tim are looking at it from a
> well-formed perspective.  The NS rec means different things for each camp.
> Let me slightly rewrite what you've written, and perhaps that will work
> better.

This pretty much sums it up. And since XML allows well-formedness, we
need to support this.

I think it also brings out a fundamental schizophrenia about XML
Schemas. Are these a type system (I'm allowed to say that -- check the
heading of this thread :) or a constraint system? Maybe it would make
more sense for XML Schemas to go hog wild on types -- since this is what
the metadata people need anyway -- and delegate constraints to something
far more versatile, such as Schematron.

-- Ron