> Actually, I think we disagree only on the point of whether
> non-qualified names in the absence of a default namespace are
> "in" a namespace. We both agree that they can be treated as
> being in a null namespace, and both theory and practice
> proceed nicely from that point.
There is no null namespace. We can say that the namespace of a name is
null, meaning that the name is in no namespace; but it is wrong to say
that the name is in the null namespace, because there is no such thing.
Of course, there is a set of names containing all names that are in no
namespace, and since a namespace is essentially a set of names (despite
the assertion of the Namespaces REC to the contrary), it would have been
easy and elegant to define this set as a namespace; but the spec chose
not to do that. Every namespace is a set of names, but not every set of
names is a namespace.
Michael Kay
Software AG
home:
work:
