Re: Question from XHTML-L: Content in CSS?

At 10:45 PM 2000-12-24 -0800, Kynn Bartlett wrote:
>This question came up on the XHTML-L mailing list; I've quoted my
>reply below which contains some context and the body of my concern.
>Namely, that by eliminating the start/value attributes for ordered
>lists and choosing to rely upon CSS 2, the XHTML working group has
>made the mistake of embedding actual -content- (the "correct" 
>number values for an item are content or meta-content) in CSS, which
>is a no-no.


>Is this something for the WCAG group (to add to our guidelines "don't
>do this even though by spec you can"), for P&F (as it represents an
>error in XHTML/HTML Strict), or even for the XHTML working group 
>itself?  I'm not sure if I'm writing to the right place or not.

This is as good a place as any to raise the issue.  The preferred place to
with it would, I think, be in PF, because we would like to use PF to head off
as many "the spec lets you do this, but don't" rules as possible before they
are needed.

[Aside, general note:  Don't worry too much about where you raise issues; just
do it.  Let the chairs worry scopes and charters - that's their job.  Might as
well give them something to do.]

However, what we are trying to accomplish in PF and GL are so intimately
intertwined that it is not easy to send it one place and feel it is being
with in the most constructive manner without at the outset setting a hook
for a
check-back from time to time.


>PS:  I think that this example is related to why I didn't like the
>      previous proposal floated on this list for resolving abuse
>      of "class."  By suggesting that semantics be embedded in
>      CSS via a property for that purpose, and presenting that to
>      the user via exposing "class" information, you're ignoring the
>      WCAG checkpoint which requires pages to be usable without
>      CSS.  What's needed is better use of the markup, not a hack
>      in CSS nor elimination of class.
>>Date: Sun, 24 Dec 2000 13:03:35 -0800
>>From: Kynn Bartlett <>
>>Subject: RE: [XHTML-L] Question on start/value attributes for <ol>, <li>
>>At 12:57 PM 12/24/2000 , wrote:
>>>I've dug around the Web looking for how to actually do this, but have not
>>>found any documentation stating how an author of an XHTML 1.0 Strict
>>>has full control *in markup* over ordered list numbering using the CSS2
>>>counter-increment and counter-reset constructs.  Maybe it's trivially easy,
>>>but the examples in
>>>are inadequate.
>>To the best of my knowledge there's no way to do this in markup,
>>only in CSS.  Which is somewhat worrisome, as the W3C's Web
>>Content Accessibility Guidelines [1] specify that pages should be
>>usable without CSS; this could prove problematic if numbering is
>>reserved for CSS.
>>Thus, the removal of start/value attributes, which are actually
>>encoded content (not presentation), is problematic.
>Kynn Bartlett  <>                   
>Sr. Engineering Project Leader, Reef-Edapta      
>Chief Technologist, Idyll Mountain Internet  
>Contributor, Special Edition Using XHTML    
>Unofficial Section 508 Checklist          

Received on Wednesday, 27 December 2000 22:10:57 UTC