Kohsuke KAWAGUCHI wrote:
> Murata Makoto wrote:
> > > Kohsuke KAWAGUCHI wrote:
> > > Another concern about the consistency is the uniqueness. That is, given
> > > a RELAX NG pattern P and a path expression E, it is desirable to have an
> > > algorithm that ensures there is at most one node that can possibly match
> > > E.
> >
> > Does this have anything to do with keys/keyRefs?
>
> I think this is a problem for any path-expression based approach.
I don't get it. Why do you think this is so important?
BTW, this is certainly doable.
Cheers,
Makoto
