RE: Thoughts on namespaces

I've read over the Japanese Proposal from Toru Takahashi and edited by
Charles Goldfarb.  Other than some differences in terminology, it is
substantially the same as my proposal.  Both have the same concept of
namespace as a means for unambiguously naming things without collision.
Both allow import and simultaneous use of multiple namespaces within an
element. Both use qualified names. Both permit URLs as identifiers of
namespaces.

The key differences are:

1.	I want to have more flexible placement of the import of a
namespace, so rather than propose addition of a new "MODULE"
construction in the DOCTYPE declaration of SGML, I use the existing
ability to add an <XML-NAMESPACE> element anywhere in the document. 

2.	The proposal allows namespace prefixes to be omitted if the
results would be unambiguous.  I proposed a more restrictive rule, that
prefixes can be omitted only if they would reference the same namespace
as one's parent element. (My purpose is to keep the namespace of an
element predictable, independent of any other namespaces that may be
introduced.)

--Andrew Layman
   AndrewL@microsoft.com