Stefan Monnier wrote:
> FWIW, there's another source of info that might be used: jitter.
> When you have two fingers pressed, not only the reported position is
> more-or-less the middle point, but it's also jittery.
> So you could detect such jitter as a tell-tale of "multitouch" and then
> use the jitter itself (orientation and magnitude) as an indicator of the
> actual position of the fingers.  Whether it would be sufficiently
> reliable to be used, I don't know.

Unfortunately, the jitter has many interpretations.
* The user is moving the fingers apart
* The user is moving the fingers together
* The user holds the finger still, but pressure varies

You may be able to get a decent multitouch with one limitation - the 
first finger stay still. With that limitation, the first finger 
establish some base point (the point to rotate around, or the point that 
stand still when something is resized.)

When the second finger comes down it does all the movement, and its 
location is at any time twice as far out from the base as the midpoint 
the FR actually register. Well, perhaps a bit less than twice as far 
out, as pressure probably is higher for the stationary finger.

Helge Hafting

Openmoko community mailing list

Reply via email to