[etoys-dev] Touch Interface thoughts

Derek O'Connell doc at doconnel.f9.co.uk
Fri Jun 24 13:11:12 EDT 2011


On 24/06/11 09:01, Steve Thomas wrote:
> How will we handle multi-touch in Etoys?

FYI: there's "touch", then 2-point fake-ish "multi-touch", then true 
"multi-point" multi-touch. Resistive touch screens are (typical) single 
touch although 2-point touch can be faked (a useful technique is looking 
for jump from left/right to middle corresponding to pressing one finger 
then another on opposite side of the screen). A surprising number of 
capacitive screens are fake-ish 2-point and shows their limitations in 
test sw that displays the actual touch points, which jump about under 
certain conditions. True (imo) multi-touch can cope with more than two 
fingers, number varies, I have one device with 10-point multi-touch.

> 2) No Mouse Fat Fingers problem:

Two problems: the area in pixels under direct contact with the 
finger/thumb, which varies depending on resolution and physical screen 
size and the almost unavoidable roll of the finger/thumb as it comes 
into contact with the surface. The former complicates how to compensate 
for small controls (eg, buttons, and assuming UI not scaled-up 
massively). Multi-sampling around the touch point for nearest/top-most 
morph helps. The roll-problem results in a small movement which when 
measured in pixels could be large and may require implementing some 
dead-band to overcome (possibly reducing the effective resolution for 
touch).

-D


More information about the etoys-dev mailing list