multi-tutch?

[Joerg Reisenweber]

Am Do  29. Mai 2008 schrieb boris.stender at gmx.de:
> Hi Joerg,
> as it seems that my mail did not reach the community list (I am not a 
subscriber, just monitoring the archives) may I ask you directly the question 
regarding "multi tutch" :-) . See below:
> 
> 
> Hi all,
> if the geometric average point is returned if multiple points are touched 
wouldn't it be possible to have a fake multi touch, e.g., by doing the 
following:
> 
> Press point A
> -> Vector A is returned
> Additionally press desired point B
> -> X = (A+B)/2 is returned
> -> calculate B = 2*X-A
> 
> Questions:
> Would this be managable?
> What time difference would be required between first and second (and 
third ...) touch, to recognize such a multi-touch action and reliable 
distinguish it from a single touch at position X?
> 
> And one more comment: If that would work you could also detect multi-touch 
gestures (even if limited), e.g., use the first point statically (maybe a 
focus centre) and handle the second point dynamically (maybe for zooming in 
and out).
> 
> 
> Boris


there are two resistors in the touchpad, one for X and one for Y. We only see 
them changing on a touch, thus giving us the coords. A concurrent second 
touch just changes the R values even more, but we get no info about whether 
we touched a second point or we moved the finger. So what you suggest is a 
gesture recognition. Things become nasty because even the "geometric middle" 
isn't true but depends on pressure of one to pressure of other touchpoint 
etc.
You really get too few info out of the device to do anything reasonable with 
it beyond singletouch (at least that's the way it is now. I plan to see 
whether we can exploit dynamic pulse response of this design to get some 
additional info. Same way you're testing cat5-cable for breaks. Don't hold 
your breath though, chances are bad. And it's not on top of my todo-list)

cheers
jOERG
-------------- next part --------------
A non-text attachment was scrubbed...
Name: not available
Type: application/pgp-signature
Size: 194 bytes
Desc: This is a digitally signed message part.
Url : http://lists.openmoko.org/pipermail/community/attachments/20080529/4ac55889/attachment.pgp