- From: Dave Raggett <dsr@w3.org>
- Date: Sat, 30 Dec 2006 11:47:25 +0000 (GMT)
- To: Anne van Kesteren <annevk@opera.com>
- Cc: James Justin Harrell <herorev@yahoo.com>, www-style@w3.org
On Sat, 30 Dec 2006, Anne van Kesteren wrote: > Well, this gets back to the original point. Why would you send > proprietary markup languages over the web... You shouldn't. The ability to render application specific XML data either via CSS or XBL is is arguably better than being forced to transcode the data into HTML or to represent it as data in a script. Another case for sending XML is where you want to provide declarative extensions to an existing standard, and intepret them via CSS and script. The extensions may later be standardized, but the flexibility to try things out before the standardization process is definitely valuable. The CSS3 selectors spec provides a means for substring matching that can act as an alternative to the HTML class attribute. It is not an exact match, since it has different specificity, and it doesn't explicitly support the notion of whitespace separated lists of tokens, but it is good enough for the purpose. Another point of flexibility is whether applications can extend CSS without having to wait for browsers to implement the extensions. The selectors API provides syntactic convenience in building a list of DOM nodes, but as far as I can see, it doesn't give developers the means to extend CSS with new properties. I think we are lacking an interface to declare new CSS properties. This would involve such information as whether they inherit or not, what the initial value is, and how to parse the value, e.g. using the syntax used in the CSS specifications. Such an interface would make it easy to experiment with proposed extensions to CSS, such as the draft CSS3 speech module. I did take a look at the Mozilla source code with a view to adding support for the CSS3 speech properties, but it proved to be a much bigger task to declare new properties than expected and something that only someone intimately familiar with the Mozilla CSS code base could manage. A lightweight declarative means to declare properties that are interpreted via the getComputedStyle interface would be really valuable and something that was overlooked in previous work. Dave Raggett <dsr@w3.org> http://www.w3.org/People/Raggett
Received on Saturday, 30 December 2006 11:47:38 UTC