- From: Ian Hickson <ian@hixie.ch>
- Date: Thu, 16 Oct 2008 21:08:30 +0000 (UTC)
- To: Daniel Glazman <daniel.glazman@disruptive-innovations.com>
- Cc: David Hyatt <hyatt@apple.com>, Matthew Raymond <mattraymond@earthlink.net>, W3C CSS <www-style@w3.org>
On Thu, 16 Oct 2008, Daniel Glazman wrote: > David Hyatt wrote: > > > > Another (IMO simpler) idea would be to just have label match the same > > pseudo-classes that the control does, i.e., if a checkbox is :checked, > > then the label can match :checked too. Same for :disabled, :enabled, > > :indeterminate, and :focus. I don't see any reason to introduce new > > selectors to solve this problem. > > Let's go back in time for a second since this issue was raised precisely > 3 years ago : my opinion as both Selectors' editor and CSS WG > co-chairman is that this issue is outside of the scope of the CSS > Working Group and it's up to the markup language to state that this or > that state set on the control is mirrored onto the label having a |for| > attribute targeting that control. To summarize, it's IMHO an HTML5 > issue. I agree that it should be up to the markup specs to define which of these state-based pseudo-classes should apply to which elements. Should we define a set of common terms so that specs can "interface" on this issue, or would it be better for specs to just name pseudo-classes explicitly? -- Ian Hickson U+1047E )\._.,--....,'``. fL http://ln.hixie.ch/ U+263A /, _.. \ _\ ;`._ ,. Things that are impossible just take longer. `._.-(,_..'--(,_..'`-.;.'
Received on Thursday, 16 October 2008 21:09:06 UTC