Playing with touch screen events
Carsten Haitzler (The Rasterman)
raster at rasterman.com
Mon Oct 20 13:10:25 CEST 2008
On Mon, 20 Oct 2008 18:00:36 +0800 Tick Chen <tick at openmoko.com> babbled:
yup. most of these i have observed just in doing the ui stuff. drags are
basically unreliabel as u get discontinuous events (down, move, move, move, up,
down, move, up, down, move, move, move, up) etc. if you press very hard its
reliable. and the offcenter press is another problem. you would hope/think the
press is in the middle of your finger, but its actually just at the very
top/bottom or just off it above/below.
what we need is a better touchscreen.. or a very very good filter between the ts
device and x events so every app gets reliable events. tslib i think has such
an infrastructure - BUT i never got time to play with it. :)
nb - kinetic scrolling. inside e it actually collects a whole series of events
(the 20 most recent mouse events) and computes motion over all of them (as long
as the events are within a given time of the current time, so very old events
are not considered. the time window is 0.2 seconds). so this tends to work well
- etk does it differently. reliable filtered ts events at a lower level will
make such filtering less important/needed.
good work there in trying to numerically summarise these things that i've
noticed too - i just knew what was happening as a bi-product of input, never
with explicit raw data to show it!.
> Hi List,
> In order to make users to have a better experence of using touchscreen,
> I wrote a small application to observe the touch screen events.
> And found some very interesting phenomenons.
> 1. If the touch pressure is very light, the status of mouse will
> becomes very unstable. Mistake move events as up and down events.
> It may need a better threshold.
> 2. It's very hard to touch the margin of Neo touch screen with fingers.
> (Therefore UI design have to be careful about that.)
> 3. When we touch a button with figer, the actual position we touched may
> have a offset to that we expected. (That means buttons should be very
> 4. start/stop scrolling may encounter unstable evnets. (I think it's
> because of the change of largest pressure position.) That means if we
> want to have a good kinetic scroll view, widget library may need to
> discard the first/last few events.
> I had record about the phenomenons 1, 2 on youtube.
> Sorry for my bad english, I am improving it. ;)
------------- Codito, ergo sum - "I code, therefore I am" --------------
The Rasterman (Carsten Haitzler) raster at rasterman.com
More information about the devel