I agree that unqualifed attributes "have no namespace" is a mistake.
Even though I did a double-take when I read that part of the spec, it
still tripped me up in my first XML application.
Unfortunately, it's not aesthetics, but the unknowably-sized matter of
deployed base that concerns me. I think the best we can hope for is to
say that the if an element is part of a namespace, then unqualified
attribute names violate validity constraints. I *might* go along with
"the result of their presence is undefined," but I think core XML has
managed to avoid such "is out of scope" issues so far, and I'd hate to
to be the first.
How's that for a compromise?
/r$
