W3C home > Mailing lists > Public > www-style@w3.org > April 2004

Re: content: url() is bad

From: David Woolley <david@djwhome.demon.co.uk>
Date: Mon, 12 Apr 2004 10:42:09 +0100 (BST)
Message-Id: <200404120942.i3C9gAI06641@djwhome.demon.co.uk>
To: www-style@w3.org

> How is this worse than list-style-image?  Or is it?

Almost nobody uses list style image, they put the image into the HTML
(and don't use list elements)!

The problem, especially, if you do not enforce a restriction to generated
content, is that most page designers treat HTML, CSS and DOM and EcmaScript
as a single language and will mix and match features from all of them
to achieve their intended visual and behavioural result, so if CSS generated
images allows a certain hack, they will use them, even if they image is
real content.  They already use CSS background images as real content,
to construct fancy controls[1], and it is standard practice to use content
to achieve purely presentational effects.

Most images in web pages aren't actually content, but styling (e.g. the
famous spacer.gif, but also fancy rules, etc.; even apparent content may
really be purely for mood, just like the choice of colour).  The question,
therefore, is what tradeoff between real content images in CSS as against
styling images in content will result, when real authors get involved;
presumably with real content in CSS getting a higher penalty that styling
masquerading as HTML content.

This actually comes down to the "complexity" thread, running on www-html;
the more features you add, the more ways there are of abusing them.  However,
only when real world authors are let loose will you find out which ones are
really abused.

[1] The site that originally caused me problems with this - the hot spot
in the HTML only aligned with the right part of the image in IE - not in
Mozilla, seems to have cleaned up its act on this one, though.
Received on Monday, 12 April 2004 05:59:42 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Monday, 27 April 2009 13:54:28 GMT