Namespaces and XML validation

[John Cowan]

Tim Bray wrote:

> >That concedes in effect that there are instances which
> >simply *cannot* be validated, because they use the same QNames
> >in inconsistent ways.
> 
> That doesn't follow; you can certainly construct a DTD to describe
> any conceivable well-formed instance.

In the trivial sense, yes.  You can declare all your elements with
content model ANY, and declare every attribute which actually
appears as CDATA #IMPLIED.  I believe there is a tool somewhere
that actually constructs this sort of DTD, given an instance.

> If what you're saying is
> that a single namespace contains usages of the same element or attribute
> that are so wildly inconsistent that a DTD won't be helpful, then
> that is a problem of that namespace which would exist even were it
> standing alone - thus is orthogonal to the issue of namespaces. -Tim

No, not at all.  I am saying that elements (or attributes) may share
QNames but come from different namespaces and have different
content models (or attribute types).  That is a condition which couldn't
exist under the old draft.

Since DTDs know only QNames, they can't cope with such conflicts,
not between colliding namespaces, but between colliding prefixes.
In other words, if prefixes were always mapped 1-1 to namespaces,
this problem wouldn't exist.

This is quite separate from locality *as such*; locality of prefixes
just means that a given prefix isn't legal outside the scope of
its declaration.  However, following the Algol precedents, you have
allowed declarations to supersede conflicting outer declarations
and to coexist with conflicting non-overlapping declarations.
There is some, if not overwhelming, software engineering experience
to indicate that this is a Bad Idea.

In short, this draft is essentially SUBDOC all over again: the scope of a namespace
declaration has fundamentally different element and attribute
types from its surrounding context.

-- 
John Cowan	http://www.ccil.org/~cowan		cowan at ccil.org
	You tollerday donsk?  N.  You tolkatiff scowegian?  Nn.
	You spigotty anglease?  Nnn.  You phonio saxo?  Nnnn.
		Clear all so!  'Tis a Jute.... (Finnegans Wake 16.5)

xml-dev: A list for W3C XML Developers. To post, mailto:xml-dev at ic.ac.uk
Archived as: http://www.lists.ic.ac.uk/hypermail/xml-dev/
To (un)subscribe, mailto:majordomo at ic.ac.uk the following message;
(un)subscribe xml-dev
To subscribe to the digests, mailto:majordomo at ic.ac.uk the following message;
subscribe xml-dev-digest
List coordinator, Henry Rzepa (mailto:rzepa at ic.ac.uk)