xlink's usage of prefix qualified attributes

Hi all,

I'm currently working on a project which may use XLink. In general, I find
it quite nice; However, one thing I find quite frustrating is that to
process it, it requires one to use namespace-qualified attributes
(xlink:label, xlink:title, etc.). Although I understand that going forward,
there will be increasing usage of namespace-qualified attributes (especially
with XML schema-aware tools), it has significant problems from an existing
tool basis. First, I'd like to outline the challenges I see, and second I'm
wondering if people have considered possibilities to this:

1) To start off with, one must consider what it means to have
namespace-qualified attributes. The namespace specification declares the
"partitions" that namespace-qualified nodes (elements and attributes) should
be used for, in Appendix A.2 of the "XML Namespace Partitions" document. It
clearly says that in order to have namespace-qualified attributes, then
these must be "global attributes". The exact wording is:
	In XML documents conforming to this specification, the names of all
qualified (prefixed) attributes are assigned to the global attribute
partition, and the names of all unqualified attributes are assigned to the
appropriate per-element-type partition.
Now, this is really problematic. DTDs--for all intensive purposes--have no
concept of "global" attributes. Most early schema languages (SOX, XDR, etc.)
don't support global attributes either. Thus, if one wants to declare
schemas with global attributes, one must define the schema in XML schema,
which is a problem for the majority of tools.
2) In addition, even if one can create a schema that uses XLink attributes,
most tools today which create instances (other than very low-level tools
such as syntactic XML editors) aren't very good at generating instance
documents with namespace-qualified attributes. A good example of this is the
Tibco Extensibility XML toolset, which IMHO are some of the best XML
authoring tools out there. However, consider the case when I want to write a
schema adjunct in their Turbo XML toolkit. Well, it turns out that when one
uses prefix-qualified attributes, it doesn't even declare the namespace
declaration for that prefix, thus creating non-wellformed XML. If one looks
at legacy non-XML to XML translators, one also runs into similar problems.
Taking these two considerations together, I was wondering if you considered
ways to create xlink-compatible documents that did NOT require
namespace-qualified attributes? 
Kelly Schwarzhoff