Last call comments XPointer Framework/xmlns()/element()

Excellent drafts. Thanks.

FRAMEWORK

2 Conformance.
Is it a requirement that an XPointer identify a subresource? (Which is
suggested by 3.3)

Case in point. We (the HTML WG) are producing XFrames as a replacement to
Frames.
(http://www.w3.org/MarkUp/Group/2002/WD-xframes-20020531)

That draft uses a query-style URL to populate the frames of a frameset.
Since this breaks caching, we are being asked to consider another,
fragment-based style instead.

If we used
    home.frm#frames(main=main.xml,nav=nav.xml,rest=blank.xml)
is this an XPointer, even though it doesn't "identify a subresource"?

3.2 Shorthand pointer
I realise that this is not exactly your bailiwick, but can't you push for
the introduction of an xml:id attribute that can be used throughout XML that
will work equally well for validating as non-validating UAs?

3.3 scheme-based pointer
Please give a (non-normative) example or two to help the poor reader!

3.4 Namespace binding context
This is very hard to understand if you haven't already read the xmlns()
scheme spec. An example would give relief!

XMLNS

I have a slight aversion to the use of 'foo' and 'bar' in examples, and
would prefer something more real world. How about using SVG's and XHTML's
<a> elements:

    <doc xmlns:xlink="http://www.w3.org/1999/xlink">
        <svg:a xmlns:svg="http://www.w3.org/2000/svg xlink:href="..""
            <xhtml:a xmlns:xhtml="http://www.w3.org/1999/xhtml"
href="..">This element and its parent are in different
namespaces</xhtml:a></svg:a>
    </doc>

(or would this be too painful :-)
Do as you wish. This isn't a showstopper for me.

ELEMENT

Replace "foo" with "intro".

Best wishes,

Steven Pemberton
Chair, HTML WG

Received on Wednesday, 31 July 2002 09:58:35 UTC