[
Lists Home |
Date Index |
Thread Index
]
"K. Ari Krupnikov" wrote:
>
> james anderson <james.anderson@setf.de> writes:
>
> > "K. Ari Krupnikov" wrote:
> > >
> > > Arjun Ray <aray@nyct.net> writes:
> > >
> > > ...
> > >
> > > So instead of /a:foo/b:bar, you'd write
> > > /foo[@xmlns='a-role']/bar[@xmlns='b-role']
> > >
> > > Interesting.
> > >
> >
> > i believe the complete proposal includes (or at least suggests) a dom access
> > layer which effects the relabeling and permits
> >
> > /foo/bar
>
> Clearly, this XPath matches a superset of nodes matched by the one
> above.
>
not necessarily.
one way to think of xmlns-based relabeling is that it translates label-sets
(conceived in a two-d plane) from a set of variably located lines specified by
the prefixes to a single constant line specified by the namespace name.
one way to think of attribute-based relabeling is that it translates label
sets from a constant origin line to a set of variably located lines specified
by the attribute values.
the size of the respective sets of labels does not change.
...
|
