Re: About dropping the style attribute

On 6/27/07, Karl Dubost <karl@w3.org> wrote:
>
>
>
> Le 27 juin 2007 à 14:49, Daniel Glazman a écrit :
> > Entirely agreed. Instead of REMOVING the style attribute, why don't we
> > discuss EXTENDING it to full CSS styles and not only declarations,
> > taking advantage from the CSS error handling rules ? I remind you that
> > the CSS WG has a document about it.
>
> For reference
> http://www.w3.org/TR/css-style-attr


Are we speifically NOT discussing that draft because it belongs to another
WG?  Should we involve that WG in this discussion before creating a new
element that does exactly the same thing with different syntax?  Has the CSS
WG officially declared that draft dead?  It seems odd to me for this WG to
drop that style attribute in spec that published around the same time that
another WG published a spec that defines the style attribute's syntax.

There is at least one reason that @style is better than HTML4 presentational
attributes: its specificity is well defined in the cascading order of CSS.
Where does @width or @border fall in that specificity does #someid { border:
0; } override it?  what about #someid { border: 0 important!; } ?  @style is
a better choice in that case for that reason.  As far as placeholder boxes
as a page loads: @width or @height works exactly as well as @style.

Whatever alternative is proposed for @style must be just as convenient to
authors in HTML and in scripting.  If it's more convenient to create a
@style attribute than it is to assign an id to a unique element and create a
<style scoped> element earlier in the document, authors will do the former.
Even if @style is nonconforming, if it works (it will) and is more
convenient than the conforming alternative, authors will use it.

-- 
Jon Barnett