W3C home > Mailing lists > Public > www-style@w3.org > January 2012

Re: CSS Referencing

From: Matthew Wilcox <elvendil@gmail.com>
Date: Thu, 5 Jan 2012 17:08:56 +0000
Message-ID: <CAMCRKiKxSPAZ1NK8hR391Fis6csr0F_7_BSSqBPxSE+Q-6cyzA@mail.gmail.com>
To: "Oliver St. John" <osj@cbord.com>
Cc: "www-style@w3.org" <www-style@w3.org>
I think it's so powerful because it stops styling relying on the
mark-up structure. If you can reference outside of the current element
and it's immediate ancestors you're no longer trapped with having to
author exactly nested mark-up to meet a style requirement. That, to
me, make it worth serious effort/consideration. Our reliance on the
cascade alone is the biggest flaw we have in CSS, in my opinion. (The
Cascade is essential and genius, but it's not enough to allow true
separation of the web-design layers jobs).

We'd have to think about sensible strategy as you say.

For example, elements that reference each other should go with the
largest shared value: e.g.,

div#one { height: #two; }
div#two { height: #one; }

would ensure equal height columns, but to whichever is the greatest
height after content is inserted.

That's a rather simple syntax and might be limiting, so perhaps a
construct like:


might work well. E.g.,

div#one { height: get('#two', width); } //sets div#one height to div#two width.

You could even combine with Calc:

div#one { height: calc(get('#two', width) / 2); } // makes div#one
half the width of div#two

For layout, I think it'd be incredible to be able to reference another
elements property. e.g., move div#one 20px offset from div#two,
regardless of the mark-up structure:

div#one { position: #two; left: 20px; top: 20px; }
Received on Thursday, 5 January 2012 17:09:29 UTC

This archive was generated by hypermail 2.4.0 : Friday, 25 March 2022 10:08:09 UTC