[whatwg] Suggestion: New "input-redirect" CSS property.

[Sander]

Ian Hickson wrote:
> On Tue, 17 Aug 2004, Matthew Raymond wrote:
> > 1) The new behavior is not backwards compatibility with HTML 4.01, thus
> > resulting in unexpected behavior when webmasters who wish to start
> > using our new markup.
> 
> I think this is a specious argument. I have yet to meet an author
> (except, apparently, you) who is using <label> because they specifically 
> want focus transfer even when the user's platform wouldn't have it.

*waves hand* And me.

Well, for some values of "want". With the new paragraph on accesskey, I'm
happy enough with the current proposal, but in the past I have purposefully
used label to take benefit of the focus transfer (though it was also the
semantically right thing to do).
Specifically, I had to deal with a design (which was a given - I had no
power to change it) with a "subscribe" input (email) with a header/label
sitting directly underneath a column of navigational links. This header and
the navigational links were styled nearly identical. The navigational links
were of course meant to be clicked - but the header for the subscription
box wasn't? So it became a label. It still sucks, as the visibility of the
cursor isn't very high, but at least there's now _something_ happening to
clue in the user why his click didn't produce the expected effect.

(Is this description clear? I unfortunately can't link as the site isn't
live yet.)

Anyhoo, just so as to show that situations like this do exist in the real
world.

> I've never ever seen anyone use <label> with <object> and would be
> tempted to say, in the WF2 draft, that using <object> as a form control 
> is no longer allowed. In practice I'm not aware of anyone who has ever 
> actually done this. The NSPAPI doesn't, to my knowledge, even support it.

It probably isn't done because it doesn't work. (chicken, egg) Yet I _have_
written a few quickly thrown together apps (purely for personal use) where
it's not too much of a stretch to envision me using this. Think things in
the realm of data validation/extraction from multiple databases/systems,
where access to at least one would be through a java applet. For
efficiency, you want this to be purely keyboard based, using accesskeys to
quickly focus the correct input fields and submit - and then ideally also
using an accesskey to move focus to the applet to access the data in
_there_ (I've never done exactly this, but I have done the same with an
iframe. Very annoying to have to reach back to the mouse each time I wanted
to get something from in there. If it had been worthwhile timewise, I'd
have gone to write some script to catch keypresses and manually move focus,
but unfortunately it wasn't.)

> > 7) As it stands now, <label> is underused on the web. Many webmasters 
> > use text only. Removing focus passing to the control will only
> > encourage this trend.
> 
> I don't see why. The reason to use <label> has nothing to do with focus 
> passing, it has to do with semantic markup, aiding accessibility tools, 
> and so forth.

Focus passing is probably the _one_ 'reason' which has any effect at all to
convince the many clueless authors out there to start using labels.
_Showing_ the benefits in doing things the right way and then later
explaining the reasonining behind it is a far more effective way to get
through to people then doing it the other way 'round.

Of course, I don't really particularly care about other authors using labels
- I usually get to them through larger things, but still. :)

Sander