Re: IDs, Classes : could CSS usefully use a third (abstract) concept ?

> and that these
> proposals tend to assume that there is no cascade and only a single
> author and fail to consider how the new feature will cascade.


I don't think a proposal such as @define {}[1] ignores any of these 
issues. Having a syntax similar to entities in SGML and XML would be 
very helpful for CSS developers. I also think the argument against such 
a syntax

>in a well-written CSS there should be relatively little
>duplication of values ... so if you are only using
>constants here it may be wise to re-examine the CSS entirely.
>
Is invalid. Often I use color schemes in a web site. The color of the 
borders for one element is the background for another, and the color of 
linked anchors in a third. Being able to specify

div { border-color: &blue\; }
a[href] { color: &blue\;; }
p { background: &blue\;; }

Would be very helpful for maintenance /especially/ when developing with 
multiple authors and a large cascade. During a recent redesign project, 
we had  a printout of the site we were working on with color names typed 
on top so that our Site-wide CSS, Graphic and page-specific CSS 
developers could all be on the same page.

Also a syntax like @define could allow great interaction with imported 
stylesheets. If the format were strict such that the  CSS document  
would have to be structured  as

- @definitions
- @imports
- css rules

then perhaps definitions in the originating files could be applied in 
the imported files as well, allowing for a very slim way for sites to 
achieve different color schemes for each section.

Also, the failure of the argument for using server-side technology is 
that if the technique is documented in CSS, than there is much less 
chance for it to be mis-used, especially in the case of sending HTTP 
headers, and the differences between different server-side technologies. 
Besides, couldn't the same be said for @import?

That being said, I think the crux of this issue as that the CSSWG has 
said "no," and doesn't seem very open to comments here, regardless of 
its utility. And, I admit, for good reason. The existing spec is large 
enough not to be fully adopted my mainstream browsers (and therefore 
developers) for the next ten years. I think the argument that such 
functionality is well beyond the scope of CSS 3 is the best stance to 
take, but we would be fools to omit considering something of the like 
for later versions.

[1] http://lists.w3.org/Archives/Public/w3c-wai-ig/2002OctDec/0414.html
-- 

Ryan Cannon
Instructional Technology
Web Design
RyanCannon.com <http://ryancannon.com/?refer=email>
(989) 463-7060