Re: target-attr() and /reference combinator/ proposals

>    A /attribute/ B
> Is anyone violently disgusted by this combinator proposal?
At least there shouldn't be a space between A and /attribute/

A /attribute/ B
means an element B with an ancestor that (1) is pointed to by *any* element
and (2) has an ancestor A

An alternative could be
A:follow(attribute) B
But this is just another misuse of the colon.

>    A:selected /href/ [title] { color: green; } /* CSS3 WD syntax */
This should then be A:selected/href/[title]

> The above would match an element A only if the element which it points
> to (with its "href" attribute) has a title attribute. (I'll note again
> in passing that my :matches proposal is not particularly longer than
> the WD's version of the same idea, except that it _is_ a selector,
> unlike the 'imposter' known as :selected, and that it is much easier
> to extend in later CSS versions to allow more complicated expressions.)
(I'll note again that my proposal, ::pseudo-element and :: as 'self',
is much shorter:
A/href/[title]:: { color: green; }
^_^ )

> Does it seem reasonably feasible to implement?
I think so.

Sjoerd Visscher