Re: [css3-values] attr() function

On 8/2/05, Bjoern Hoehrmann <derhoermi@gmx.net> wrote:
> * Orion Adrian wrote:
> >I think it's unreasonable to think that languages will always use the
> >CSS color syntax. There will always need to be some sort of conversion
> >function. With HTML4, it's fairly simple, but even so, there's no
> >conversion function even for something as simple as putting a # in
> >front.
> 
> This is true, X3D for example uses a list of float for colors, for
> Avalon objects in XAML you can use #rrggbbaa, other languages like
> SVG don't even share the same model (you'd specify the fill color
> and the fill opacity separately); for other data types there are
> similar issues, resource identifiers in attributes typically don't
> use the url() functionality notation, float values might allow for
> a 'scientific notation' and so on.
> 
> I don't see how this matters here though, the cases where you'd want
> to use styling attributes and CSS for the same document and actually
> can because the styling model is sufficiently similar seem uncommon.
> Some features in CSS require attr() and it has proven useful for
> e.g. mapping legacy styling attributes like those in HTML to CSS in
> some user agents, and making it more general purpose does not add
> significant overhead to CSS implementations; but it's not meant to
> be an ubiquitous feature.

Doesn't this constitute a lost opportunity? So CSS is supposed to
format SMGL style languages that may encode data a certain way and use
attributes and not child elements for storing stylisitic data? Isn't
that just another word for HTML and XHTML?

How about we call CSS for what it is, the styling side of XML/HTML and
be done with it. If we admit that, we at least get to design them in
tandem with each other in mind. Not admitting that doesn't get us
anywhere since CSS can't effectively be used outside SMGL
applications.

-- 

Orion Adrian

Received on Tuesday, 2 August 2005 14:03:18 UTC