Re: [css3] "Selectors that People Actually Use"

Jens Meiert:
>
>>> a few CSS 3 selectors might indeed require a revision.
>
> * E:empty -- Chances are indeed that authors will use element selectors
> instead,

It might be hard for many authors to distinguish elements that are empty,
replaced, contain no text (only or not even whitespace) or no children (only
or not even whitespace). I do not know whether there is a use case for each of
these.
The specification should probably include these examples:

  <foo> </foo>
  <foo/>
  <foo src="bar"/>
  <foo><bar/></foo>

  foo {content: "bar";}

Btw., an intentionally blank section -- what the hell?

> * E:lang(fr) -- Probably the most questionable selector since apparently
> replaceable by attribute selectors.

I still think it was too strict to make it impossible for this pseudo class to
be used with markup languages that do not use IETF-compatible language tags.
It was the right choice to make the selector syntax itself compliant, of
course.


It has been said several times in this thread that Selectors is intended to
work with more than HTML and interactive user agents. It should indeed, yet
some proposals, e.g. using certain characters like '<' and '&' or requiring
reverse lookup and thereby making matching slower, have been rejected on the
very presupposition that the spec must work primarily with HTML in browsers.

Perhaps there should be profiles or inline notes to tell implementors in which
environments certain selectors have to be supported to meet compliance and
where they are optional or even forbidden.

Received on Tuesday, 19 February 2008 15:56:07 UTC