Issues with HTML fragment [was XHTML Requirements and Timing] (fwd)

The XML Linking working group is concerned over the way id and name
are defined on the <a> element. Daniel Veillard's suggestion is that
the HTML spec specifies name as ID and id as an nmtoken for <a>.
Note existing html browsers tend not to recognize the id attribute
so this change would bring XHTML into closer alignment with the
current browser base.

Regards,

-- Dave Raggett <dsr@w3.org> http://www.w3.org/People/Raggett
phone: +44 122 578 2984 (or 2521) +44 385 320 444 (gsm mobile)
World Wide Web Consortium (on assignment from HP Labs)

---------- Forwarded message ----------
Date: Fri, 16 Apr 1999 13:57:26 -0400
From: Daniel Veillard <Daniel.Veillard@w3.org>
Subject: Issues with HTML fragment [was XHTML Requirements and Timing]

> http://lists.w3.org/Archives/Member/w3c-xml-cg/1999Apr/0057.html
> 
> The XML linking WG will review subsequent "last calls" for XHTML with
> the
> proviso that we will only look at those sections where XHTML and XML
> Linking overlap *and* changes have been made. A commitment to a more
> thorough review would necessitate further delays in our deliverables.
> Further, the timing of the delivery of the "package" presents additional
> difficulties for this WG since we are participating in a joint effort
> with
> the XSL WG and their deliverable dates (as we understand them) have
> considerable overlap with the XHTML dates. The final complication for
> some
> WG participants is our travel schedules over the next (at least) 4
> weeks.
> 
> We're out of resources and out of cycles. If XHTML needs more thorough
> review by this WG, it needs to be delayed.

 We (XML Linking WG) have a serious problem with the current version of the
spec, we cannot accept name attributes to be of type CDATA and continue
to support the HTML fragment notation. Basically with the current version
of the xhtml DTDs #xptr(foo) point to the first A element carrying an
name attribute of value "xptr(foo)". This leave absolutely no room for the
Linking WG to preserve backward compatibility with HTML fragment scheme.
 We need the fragment values to be restricted to Name space to be able
to build a decent scheme for handling the ambiguity. Failure to do so
will lead to major incompatibilities on whether an xhtml document is
provided with his own mime type, mimetype text/xml or application/xml

 We suggest to upgrade name attribute to be of type ID. This currently
poses a problem w.r.t. the XML specification which (without good reasons)
restrict the number of ID type attributes on an element to 1 in the DTDs.
 Hence since id is declared of TYPE ID everywhere we have a problem. The
long-term solution is a modification of the XML spec, Tim Bray is nearly
100% sure that this is not a problem, and this is a suggested item for
the XML cleanup scheduled in the XML Plenary meeting. The short term
solution is more hairy, one could for example downgrade id to be just
NMTOKEN on the A element (and the 8 to 10 other elements holding name).

  My personal opinion is that this name/id stuff is a fiasco, HTML clients
tend to not support references to id, and this induce problems due to
overloading, for example when id="foo" and name="foo" are both present
on a document and not on the same elements. The suggestion to use
both id="foo" and name="foo" on xhtml (on the non-normative part though)
wasn't considered a good thing by the Linking working group. The proper
way to do this is to get name to be of the proper XML attribute type.

Daniel
-- 
	    [Yes, I have moved back to France !]
Daniel.Veillard@w3.org | W3C, INRIA Rhone-Alpes  | Today's Bookmarks :
Tel : +33 476 615 257  | 655, avenue de l'Europe | Linux, WWW, rpmfind,
Fax : +33 476 615 207  | 38330 Montbonnot FRANCE | rpm2html, XML,
http://www.w3.org/People/W3Cpeople.html#Veillard | badminton, and Kaffe.

Received on Monday, 19 April 1999 11:27:30 UTC