For example: inability to select the parent element from the child rule is one of the most basic functionalities in CSS which is currently missing.
Let's argue.
While a parent selector would be extremely useful, it has some performance issues that are apparently hard to solve with rendering engines. It's not been left out due to some kind of oversight by the W3C, it's been debated quite a bit at WG meetings. Much the same as element-level media queries (width of parent element).
I think selectors (and combinators and pseudos) in general have become pretty powerful. eg. I have a Selenium test for table sorting which just uses the nth-child selector. It's certainly better than the days when you could only really use element, class or ID selectors.
Why would you use/need that capability, in your example? Just curious :)
estevan carlos benson
Artist & Technologist
Link to specs?
Also, I think it's reasonable to not allow a child element to select a parent. I'm under the impression it would require additional resources in order to read through the markup an additional time for the sake of finding the parent outside the child.