Touch-pad expectations

[Paul Fox]

hal wrote:
 > 
 > richard at laptop.org said:
 > > One absolute we know is that the touchpad hates multi-touches.  If you  have
 > > more than 1 point of contact with the sensor then it will go  bonkers.  Due
 > > to the button placement below the touchpad this is quite  easy to do.  Its
 > > trivial to show that double touches throw the pad into  a tailspin.  Other
 > > failure modes are harder to duplicate. 
 > 
 > Thanks.  That may be part of my problem.
 > 
 > I keep my left thumb on the left button.  My left hand is underneath the XO 
 > and the thumb wraps around to the button but doesn't get near the touchpad.  
 > But my ring finger may be drooping down and getting close enough to cause 
 > troubles.  I'll try to keep an eye on that.
 > 
 > 
 > I would really like to see some feedback when it is doing a recalibrate.  Can 
 > somebody point me at the right section of code and/or provide a few hints on 
 > how to implement that?  I'm thinking of a hack, nothing clean and elegant.  
 > For an experiment, I'd be happy to dedicate any corner of the screen.  (Say 
 > 1/4 inch sq, or bigger if I can't see that fast enough.)

the recalibration happens in the touchpad driver:
    http://dev.laptop.org/git/olpc-2.6/tree/drivers/input/mouse/hgpk.c?h=olpc-2.6.31#n601

there's no direct indication to userspace when a recalibration occurs,
but during development i had a script of some sort that watched for
the kernel message you see there, i.e. "Recalibrating touchpad..", and
beeped or something when it happened.

there are other tuneables available as well -- feel free to play
with them, obviously.  see here, and following:
    http://dev.laptop.org/git/olpc-2.6/tree/drivers/input/mouse/hgpk.c?h=olpc-2.6.31#n78

paul
=---------------------
 paul fox, pgf at laptop.org