[XML-SIG] ANN: XIST 2.0

Daniel Veillard veillard@redhat.com
Thu, 17 Oct 2002 13:02:38 -0400


On Thu, Oct 17, 2002 at 06:43:08PM +0200, Walter D=F6rwald wrote:
> > Please
> > read http://www.w3.org/TR/REC-xml-names/=20
>=20
> Ouch, seems I missed the "No entity names, PI targets, or notation
> names contain any colons." part on the first reading.

  Sorry about that. If you actually go back far far in history
you would see that initially namespaces were declared with PIs but
this has been dropped.

> > ----
> > <?foo:bar?>
> > <doc/>
> > ----
> > <doc>
> > <?foo:bar?>
> > </doc>
> > ----
> >=20
> >   first case there cannot be any namespace in scope.
> >   second case is perfectly legal w.r.t. XML + Namespace and still you=
r
> >   mapping has no semantic.
>=20
> Yes, but
>=20
> <foo:doc>
> <?foo:bar?>
> </foo:doc>
>=20
> doesn't either.

  Right,

> <doc xmlns:foo=3D"http://www.foo.org/foo">
> <?foo:bar?>
> </doc>
>=20
> however does.

  No neither, The semantic of the namespace applies only to elements and
attributes, and again <?foo:bar?> is not correct w.r.t. XML Namespace.

> So is there any better way to fix this problem? The point of XIST

  Namespace are not intended to cover PIs. Don't try to force the
spec in a behaviour  or semantic it clearly wasn't intended for.
Make your framework implement XML to provide higher level abstraction
but don't try to tweak the underlying specification to adapt
to those abstractions.
  Namespace applies only to element and attributes, and possibly not=20
all of them, you're operating on a subset of the nodes present in your
tree, make your abstraction follow this limitation.

Daniel

--=20
Daniel Veillard      | Red Hat Network https://rhn.redhat.com/
veillard@redhat.com  | libxml GNOME XML XSLT toolkit  http://xmlsoft.org/
http://veillard.com/ | Rpmfind RPM search engine http://rpmfind.net/