Re: Using XPointer with HTML

On Thu, 11 Apr 2002, Jim Ley wrote:

   "Charles McCathieNevile" <charles@w3.org>

  > Well, it depends what you want to do with the XHTML. Many people want
  stuff
  > that renders in a browser, and XHTML written according to the
  compatibility
  > guidelines can render in Lynx, Amaya, Mozilla beta IE on various
  platforms, iCab, etc.

  but not IE on all platforms, and not the myriad of other browsers,
  including ones in embedded systems which aren't (easily) upgradeable.
  Suggesting XHTML 1.0 with some compatibility guidlines will make your
  content render appropriately in more browsers is clear bunkum.

At the moment we are talking in hand-waving generalities - I guess we would
do well to get more clear facts. (I am not saying that what you say is
untrue, but I am intersted in looking in some detail at the information.

  HTML 4.01
  has all the accessibility features of XHTML 1.0 and has the bonus of
  being supported in a lot more browsers.  If XHTML is appropriate for
  future browsers, then the simple matter of content-negotiation on the
  XHTML mime-type is appropriate.

Yes, I think this is the key point. With a server like Jigsaw I think it is
now possible to do something smart by  validating content, and being able to
serve one file as two different content-types. (XHTML should really be sent
as xhtml+xml, but even browsers that do grok XHTML don't always know what
content-type to expect.)

  > (supports SVG plugin on
  > Mac OS, although when it gets out of beta it is expected to support it
  > natively),

  I fear you've not been keeping up with Bugzilla...

yep, true.

  http://bugzilla.mozilla.org/show_bug.cgi?id=133567
  which says it applies to ALL OS's.  I also understand there is no chance
  of native SVG making it into Mozilla 1.0, which has already been pretty
  much fixed (a Release Candidate is due on the 15th).

  > The problem is that there isn't any very well defined way to point to
  > arbitrary parts of HTML, and particularly for invalid HTML.

  The SGML normalisation is every bit as appropriate as an xhtml one, it's
  also simpler to implement as it's what the browsers and validators are
  already doing...

well, some browsers are. Some browsers are doing XML-based parsing, and don't
handle SGML - as I understand it this is the minority case which is growing.
And some browsers just have special magic processing for HTML...

cheers

Charles

Received on Thursday, 11 April 2002 08:12:11 UTC