The description of the XPath Level 1 expression language contains the
following:

"NOTE: If the element name is unqualified, i.e. appears without a namespace
prefix, then the element name MUST be matched against a matching element name
in the resource document, regardless of namespace bindings that are in effect,
including default bindings. This allows implementations to simply match element
names in the majority of cases. If namespace bindings are significant for all
elements, then qualified names MUST be used."

Oracle review comment: "this is asking for so much trouble. what if 2 elements
in different namespaces have the same local name? haven't we learned anything
from the failure that dos/windows case-insensitivity is?"

It's unclear whether the "If namespace bindings are significant . . ." sentence
applies to this case. In any case, requiring the client to know whether or not
there are multiple elements in different namespaces with the same local name
*before* they construct their expression creates a tighter coupling between the
client and the exact resource representation than we seem to have in mind.