W3C home > Mailing lists > Public > www-html@w3.org > April 2003

Re: [XHTML2] Suggestion: generalize CAPTION element

From: Masayasu Ishikawa <mimasa@w3.org>
Date: Thu, 03 Apr 2003 15:35:58 +0900 (JST)
Message-Id: <20030403.153558.71114307.mimasa@w3.org>
To: bert@w3.org
Cc: www-html@w3.org

[ personal opinion ]

> Bert Bos:
> >
> > It seems to me that associating a caption with an image is such a
> > fundamental operation, that it should be expressed by semantic
> > mark-up, rather than by visual proximity (or other purely stylistic
> > means).

You mean, not just a child but as the first child?  Do you want it
mandatory (when 'src' is used), and do you want to allow only one
caption or multiple captions?  As Christoph pointed out, the ojbect
element does something like this, although

"Christoph Päper" <christoph.paeper@tu-clausthal.de> wrote:

> I agree that this is one of the most missed features of current HTML.
> 
> The minimal content model for the 'object' element is
> 
>   (caption?, (PCDATA | Flow | param)*),
> 
> thus it's already possible--in the current draft--to caption an image with
> the prefered element for image inclusion.

This cannot be expressed in DTD, so if we generalize this to nearly
all elements, it would completely mess up mixed content model.  XML
Schema and RELAX NG can handle this better, though.  Attribute-
sensitive content model should be possible with RELAX NG, although
it may not be trivial.

Micah's "elemental XHTML" idea is interesting, though going there
with DTD would produce more side-effects than benefits, IMHO.  I'm
wondering if people are prepared to abandon DTD at this point, at
least as a normative schema.  If that's the case, we could consider
a better approach.

Regards,
-- 
Masayasu Ishikawa / mimasa@w3.org
W3C - World Wide Web Consortium
Received on Thursday, 3 April 2003 01:36:01 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Tuesday, 27 March 2012 18:15:54 GMT