Re: TAG Comments on XHTML 2.0 and HLink

> Recently, the TAG considered the scope of XLink[1]. We concluded[2] that
> XLink should be used for hypertext references in user-interface
> oriented applications.
>
> In light of this conclusion, it is the unanimous opinion of the TAG
> that XLink should be used for hypertext references in XHTML 2.0.

It is not clear you have the authority to say that. The W3C AC voted on the
issue when XLink went to Rec:

http://lists.w3.org/Archives/Member/w3c-ac-members/2000OctDec/0036.html
"b) There has been an ongoing request from the HTML Working Group
that a processor should, by reading an XHTML schema, be able to
deduce the XLink semantics of XHTML links. The W3C Note "XLink
Markup Name Control" [5] proposes two solutions. These
demonstrate that a processor aware of certain new conventions
would indeed deduce these semantics.

  [Question 2, about minority opinion (b)]

  We encourage you, in your review comments, to indicate whether
  you prefer that W3C develop these conventions on the
  Recommendation track, or leave them as a demonstration of
  feasibility."

In the announcement of going to Rec, it was stated (note that a new spin had
been added that wasn't in the original question):
27 June 2001
http://lists.w3.org/Archives/Member/w3c-ac-members/2001AprJun/0043.html

"These two issues featured heavily in the negative reviews. [...]

2) The question of a more principled resolution of problem of the
    attribution of linking semantics as defined by this specification
    in cases where existing W3C languages cannot adopt the linking
    syntax defined by this specification was already raised during the
    Candidate Recommendation review of this specifiction.  The
    decision to advance to Proposed Recommendation explicitly chose to
    proceed without such a resolution in the specifiction, but to
    continue work on the problem.  This work has made some progress,
    and a solution now seems likely to emerge from a combination of
    W3C activities in the areas of the XML Infoset, XML Schema and the
    XML Processing Model (see below): We expect to include this area
    in the responsibilities of a rechartered Linking Working Group
    when the call for participation for this group is made."

That fact that this last didn't happen, despite the apparent expectation, is
what made the HTML WG start to investigate HLink.

> * The charter of XHTML 2.0 includes the "design goal is to use generic
>   XML technologies as much as possible". XLink is an existing, generic
>   XML technology.

Yes, but "as much as possible" was added exactly to cover XLink, because it
prevents us from doing the things we want to be able to do.

> * XHTML is also expected to combine "with other W3C specifications,
>   for instance, MathML [and] SVG". Given that MathML and SVG already
>   use XLink for hypertext references, that would seem to be precedent
>   for using XLink in XHTML.

That the TAG doesn't yet understand how XLink gets in the way of combining
with other specifications will be the source for much future discussion,
which I will reserve for a later email. The fact that SMIL doesn't use XLink
would seem to be a precedent for not using XLink in XHTML.

For the HTML WG,

Steven Pemberton
Chair

Received on Thursday, 26 September 2002 08:26:40 UTC