Re: Using XPointer with HTML

From: "Jim Ley" <jim@jibbering.com>
> > > Well of course you would, they're different documents!
> >
> > No, the same document, served with different media types. It is the
> > same
> > document, but the result of the media types means it gets parsed
> > differently.
>
> Can you have an xhtml1.1 document that is served as text/html ?

That is defined by RFC 2854. I believe you could manipulate an xhtml 1.1
document in a similar way to Appendix C of XHTML 1.0
(http://www.w3.org/TR/xhtml1/#guidelines) so that it would be acceptable to
a UA that accepts text/html. It wouldn't be XHTML 1.1 to that UA of course,
but it would still be an XHTML 1.1 document. (I'm not sure if this answers
your question).

> If not
> the point is of somewhat historical interest as we're talking about a
> subset of xhtml 1.0 documents are we not?

Are we?

I don't see the relevance of XHTML 1.1 here.

> > > However if the Pointer is
> > > xpointer(id('Moomin')) then it will happily point to the same element
> > > within the Resource (assuming the resources are appropriately
> > > authored and content negotiation on the document returned from the URI
is
> > > logical.)
> >
> > As I said, if the document has ids, the problem is solved.
>
> No it is not, by your argument
> http://jibbering.example/example#xpointer(id('Moomin')) points to a
> different fragment depending on whether the resource has an xhtml or html
> mime-type.

Aha. Apparently you misunderstood what I was trying to say. No. I didn't say
that. Ignoring the problem of permissibility of using XPointer to point into
HTML, the above would identify the same element. That was what I meant with
"if the document has ids, the problem is solved."

My point was about using child elements.

    http://jibbering.example/example#xpointer(id('Moomin'))/3/4/5

could identify different elements depending on the media type that the
document was served with. This is because the same document has different
parsings, dependent on the media type (text/html, or text/xml).

So if you gave the above pointer to someone, what they would see would
depend on the settings of their content negotiation parameters.

> perhaps rather than
> limit yourself to commenting on XPointer into HTML and the context of
> TBODY you could comment on all the issues raised - those are the ones
> that are actually concerning www-annotation and wai-er after all.

XPointer into HTML and the context of TBODY was what I was asked to comment
on, and I was fulfilling that request.

Best wishes,

Steven Pemberton

Received on Wednesday, 10 April 2002 08:36:03 UTC