ash: Add some UMA for touch-events.

ash: Add some UMA for touch/gesture-events.

Count the touchpad/touchscreen interactions and the touch-event radii values. Also count scrolls on web-pages seperately.

BUG=138846
TEST=none

Committed: https://src.chromium.org/viewvc/chrome?view=rev&revision=149669

[sadrul, 2012-07-30 19:39:58]

Some notes:
* Instead of counting 'clicks', I am measuring the tap-downs in this patch.
Counting a 'click' is tricky since it's not really well-defined (for
web-contents, the logic is in WebKit, and in aura/views, we do not really have a
notion of click).
* If there is a multi-finger touch-gesture, then it will count all the
touch-points in the gesture.

Does this look reasonable? If so, I will send to OWNERS

[Rick Byers, 2012-07-31 20:37:20]

On 2012/07/30 19:39:58, sadrul wrote:
> Some notes:
> * Instead of counting 'clicks', I am measuring the tap-downs in this patch.
> Counting a 'click' is tricky since it's not really well-defined (for
> web-contents, the logic is in WebKit, and in aura/views, we do not really have
a
> notion of click).
> * If there is a multi-finger touch-gesture, then it will count all the
> touch-points in the gesture.
> 
> Does this look reasonable? If so, I will send to OWNERS

This seems reasonable - so LGTM.

I'm concerns about our ability to analyze the radius data over time (i.e. does
more large-radius events just mean more usage?), but I don't see any easy
options for addressing that (really we want count per second or something, but I
don't think we have the infrastructure for that).  Let's chat in person
tomorrow.

[Rick Byers, 2012-07-31 20:37:36]

http://codereview.chromium.org/10834070/diff/4001/ash/touch/touch_uma.cc
File ash/touch/touch_uma.cc (right):

http://codereview.chromium.org/10834070/diff/4001/ash/touch/touch_uma.cc#newc...
ash/touch/touch_uma.cc:235: 1, 500, 100);
This seems reasonable to me.  We've seen radii > 500 on some devices with buggy
firmware, but it looks like that'll just go into the overflow bucket.

http://codereview.chromium.org/10834070/diff/4001/ash/wm/system_gesture_event...
File ash/wm/system_gesture_event_filter.cc (right):

http://codereview.chromium.org/10834070/diff/4001/ash/wm/system_gesture_event...
ash/wm/system_gesture_event_filter.cc:538:
ui::IsTouchpadEvent(event->native_event())) {
I don't think we need to restrict ourselves to touchpad events here.  If a user
has an external / bluetooth mouse, it would be good to include those events in
our metric as well.  Alternately it would be nice to report TOUCHPAD_TAP_DOWN
separately from MOUSE_TAP_DOWN but I'm not sure the distinction has much
value...

[sadrul, 2012-08-01 16:15:26]

http://codereview.chromium.org/10834070/diff/4001/ash/wm/system_gesture_event...
File ash/wm/system_gesture_event_filter.cc (right):

http://codereview.chromium.org/10834070/diff/4001/ash/wm/system_gesture_event...
ash/wm/system_gesture_event_filter.cc:538:
ui::IsTouchpadEvent(event->native_event())) {
On 2012/07/31 20:37:36, Rick Byers wrote:
> I don't think we need to restrict ourselves to touchpad events here.  If a
user
> has an external / bluetooth mouse, it would be good to include those events in
> our metric as well.  Alternately it would be nice to report TOUCHPAD_TAP_DOWN
> separately from MOUSE_TAP_DOWN but I'm not sure the distinction has much
> value...

Done (using just MOUSE_DOWN for both mouse/touchpad)

[sadrul, 2012-08-01 16:56:00]

I have updated the patch to also count scroll events on webpages separately.

[Rick Byers, 2012-08-01 21:38:38]

lgtm

[Ben Goodger (Google), 2012-08-02 15:28:08]

lgtm