Re: CSS selectors and xml:id

On Tue, 10 May 2005, Robin Berjon wrote:
> 
> It's a simple question of interoperability and ease of use. You know as I do
> that IDness in non-HTML situations has been frequently problematic in browsers
> that do XML+CSS [...]
> 
> Now we have xml:id which is a very neat addition to the XML toolset. [...]
> 
> But that can only happen if it's reliably available. Which is why I think that
> CSS requiring xml:id for implementation that can apply CSS to XML [...]

By that argument:

You know as I do that DOM support in non-HTML situations has been 
frequently problematic in browsers that do XML+CSS. We have DOM3 Core 
which is a neat addition to the XML toolset. But getting it reliably 
available is the only way to make it useful. Which is why I think that CSS 
requiring DOM3 Core support for implementations that can apply CSS to XML 
is reasonable.

Or the same thing with any other pet feature.

What you want isn't for CSS to require xml:id support (xml:id has little 
or nothing to do with CSS). What you want is a "Web browser metaspec" that 
says "here are the features that browsers must support".

But Web browsers would ignore such a spec just like they would ignore such 
a requirement in the CSS spec. They'll implement the specs they want to 
implement, and no amount of conformance criteria in random unrelated docs 
is going to change that. Only bribes and patches will.

-- 
Ian Hickson               U+1047E                )\._.,--....,'``.    fL
http://ln.hixie.ch/       U+263A                /,   _.. \   _\  ;`._ ,.
Things that are impossible just take longer.   `._.-(,_..'--(,_..'`-.;.'

Received on Monday, 9 May 2005 16:00:17 UTC