several recent mails have read more into the namespaces proposal
than is written in the paper, and have then argued against what the writer
imagined the paper to say. i know it is easy to read a paper and to think
that it implies all sorts of gigantic things.
namespaces is a very simple, minimal thing. it isn't architectures.
it isn't object orientation. it isn't a mechanism for document composition.
it isn't a new way to write DTDs. it is only what the paper says: a means to
associate a name with a controlled namespace.
best wishes,
andrew