Re: [css3-mediaqueries] Touch enabled devices

Kenneth Christiansen wrote:
> On Thu, Apr 22, 2010 at 8:29 AM, Gregers Gram Rygg
>> Why isn't "-moz-touch-enabled" or similar part of CSS3 media queries
>> (without -moz)?
>> https://developer.mozilla.org/En/CSS/Media_queries#-moz-touch-enabled
>>
>> I believe this is absolutely critical to give touch-devices a better
>> user-experience.

> I think something like (main-input-method: touch) would make more sense.
> 
> With values as |touch|keys| etc. Keys could then cover additional
> cases such as setop boxes which normally are navigatable using keys on
> a remote control.

>> Since touch is not as precise as a mouse cursor.

I think the real information we're looking for is not if the user is
using mouse, finger, stylus or keypad but the precision of pointer device.

So basically we have:

(input method: precision that an average user can achieve by default)
relative pointer (mouse, trackpad, etc): 1-5 px
stylus: 1-10 px
touch/finger: 25-50 px
keypad/remote controller: element width / unknown fixed step

How about following:

pointer-precision: <len> | step;

The user is expected to be able to hit targets of size <len> without
huge effort. The <len> should be expressed in "px", "em" or "ch" unit.
The special value 'step' is used to express an UA that navigates with
keypresses or some other non-pointer based user interface. Such
interface is expected to be able to target any focusable element
regardless of size. [Perhaps it should be mentioned here that the actual
smallest target size should be considered to be <len> squared because
target size is about area, not length.]

max-pointer-precision: <len> | step;

The user is expected to be able to hit targets of size <len> with extra
effort. Targets smaller than <len> should be considered too small for
the user to be focusable at all. This could be implemented e.g. by OS as
a feature of mouse driver where holding some special key would make
mouse movement slower and user should therefore be able to target even a
single pixel target. For an user agent that normally uses step based
navigation, there could be a mode that activates a cursor that can be
moved with single pixel steps but use of this feature requires so much
effort that user is expected to use it only if nothing else works. If
max-pointer-precision is step then the user agent does not provide a
method for cursor/pointer based targeting at all.

-- 
Mikko