Re: Towards a new CSS Validator?

Here a few comments where I've elaborated on François' suggestions:

On 11 Nov 2008, at 18:29, François REMY wrote:
> Here are my few comments :
> Needed functions :
>
> 3.       Detect unknown properties + differentiate
> ·         Known but not valid in the current scheme (border-image in  
> CSS 2)
                             - Doesn't the current validator already  
do this?
> ·         Known but proprietary (filter)
> ·         Unknown but prefixed (-ms-interpolation-mode)
                              -  It would be useful if the validator  
could determine which vendor-specific prefix related to which browser  
and display the result (i.e 'We have detected that you are using a  
property that is specific to, and can only be read by [browser-name]').
> ·         Unknown (pading)
> 4.       Show multiple version of the CSS when a hack is used (which  
> value is used by which browser ?)
                        - Can't help but think this would take quite  
of effort to compile information for all hacks used by each individual  
browser? If this feature was implemented, which browsers would we  
include in this matrix?
> 5.       Detect when a rule is not possible
> ·         Due to HTML model (a > a; p > div; .)

                              - The examples you specified are still  
both valid CSS selectors, (although the markup the selectors relate to  
would be invalid HTML). Also, your feature would be doctype dependant,  
so would be a feature specific to validating against a URI of an  
external HTML document as opposed the stylesheet itself.