In Raptor the time of `appLaunch` [1] is captured in the Homescreen's gaia_grid when we have received the event and know we will be launching the app. This currently accounts for approximately 120-150ms of latency between when we trigger the touch event and when the tap is actually registered in the Homescreen. Since we control the injection of the touch events, and we also have the ability to arbitrarily create performance markers in logcat, we should experiment to see if we can reliably create the `appLaunch` marker manually at the time of event injection. If this is successful then we will finally overcome the largest discrepancy of comparing camera-based measurements to these automated timing-based ones: the initial touch latency of starting the application. At the present the Reboot phase already generated performance markers to denote the beginning of the reboot process and capture an accurate start time [2]. In theory it should be as simple as generating the marker at the same time as we inject the event [3]. [1] https://github.com/mozilla-b2g/gaia/blob/master/shared/elements/gaia_grid/js/items/mozapp.js#L309 [2] https://github.com/mozilla-b2g/raptor/blob/master/lib/suite/reboot.js#L57 [3] https://github.com/mozilla-b2g/raptor/blob/master/lib/suite/cold-launch.js#L129

If this experiment proves successful, we can remove the `appLaunch` performance entry from gaia_grid.

As we will most likely be moving to a Marionette-based model again, this won't be possible.