RE: More controlled degrading, @supported at-rule

Benjamin D. Smedberg wrote:
> How would you detect support for vendor-specific
> properties like -moz-binding or the Ms behavior property?

One doesn't have to add a way to detect vendor-specific properties. Just use
them. When they are recognized, they override any previously declared rules
(where applicable). When they are not recognized, they are ignored and the
previously declared rules are used.

Benjamin D. Smedberg also wrote:
> I don't think that named modules is great idea. Incremental support for
> CSS has been the rule for years, and I don't think that the situation is
> going to change much in the future. Mozilla already has support for
> opacity: but does not, to the best of my knowledge, support RGBA color
> values.

The previous CSS specs where contained in a single specification document
for each version. From CSS3, XHTML and DOM3 on, the specs are divided into
modules, which makes it easier to determine whether a module is supported or
not.

Benjamin D. Smedberg also wrote:
> I am quite aware that this relies on UAs to describe themself reasonably 
> accurately. The DOM specs do the same thing with DOMImplementation.

I would like to see an at-rule for CSS documents similar to
DOMImplementation.HasFeature(): cascading based on the modules that are
supported.
And about the fact that Mozilla supports only a part of the spec:
1) Most of CSS3 is still in drafts.
2) According to CSS3, modules must be implemented completely (so Mozilla is
currently not following the Color Module spec).
This left me to think that Mozilla's support for opacity is merely a feature
rather than a spec implementation.

- Daniël Pelsmaeker

Received on Friday, 17 September 2004 22:23:02 UTC