- From: Ian Hickson <ian@hixie.ch>
- Date: Wed, 8 Dec 2004 02:05:38 +0000 (UTC)
On Tue, 9 Nov 2004, James Graham wrote: > > > > > > Yeah, that makes sense. So we're saying access keys should be > > > completed deprecated in HTML? > > I disagree and I doubt whether browser makers would sacrifice HTML 4 > features that may be important in some applications (e.g. intranet > sites) even where long-standing usability issues exist. Instead, we > should merely change the text about accesskeys to note that: > > 1) UAs _should_ provide a mechanism for invoking document accesskeys > that does not conflict with OS/application keybindings (this typically > precludes alt + key or ctrl + key). The problem is that on desktop platforms, pretty much every sane modifier key is taken, leaving only pretty silly interfaces (like a modal UI where you have to switch to "access key mode", or a prefix UI where you have to hit a key before the access key); and on non-desktop UIs, the access keys specified will most likely not be available in the first place. > 2) UAs _should_ allow accesskeys to be assigned to elements where no > author-defined accesskey exists and _should_ allow author-set > keybindings to be overridden That's entirely up to the UA. Google's robot (to take one UA at random) obviously wouldn't want to assign shortcuts. > 3) UAs _should_ provide a list of accesskeys defined on the current page > (2 above will necessarily require this) Again, that's up to the UA. On Mac there is no real concept of access keys, so it makes no sense to show a list of them. > Additionally we could standardise accesskeys for authors (all with > _should_ or even _may_ ) e.g. acceskey s _should_ be used to focus a > search field. Accesskey 0 _should_ return the user to the site homepage > etc. This doesn't scale. It might work for very basic sites, but anything more complex (say, Bugzilla, GMail, or Voidwars) and it becomes useless. -- Ian Hickson U+1047E )\._.,--....,'``. fL http://ln.hixie.ch/ U+263A /, _.. \ _\ ;`._ ,. Things that are impossible just take longer. `._.-(,_..'--(,_..'`-.;.'
Received on Tuesday, 7 December 2004 18:05:38 UTC