(from discussion of logging hesitations): ... I'm glad it works as expected. I think I was looking at an older log where GESTURE triggered only once -- the log-only-once bug that you fixed.
Here are my responses you requested:
@JackMostow: I'm logging gesture from here. Or the previous method would suffice? Or both?!
-- Without knowing the difference in what they log, I'd guess that either would suffice as long as the gesture gets logged.
By doing it this way:
+: We can have Current Tutor Id & Current Student Id
Student id is constant throughout a session.
Tutor id would be useful to have but can if necessary be inferred later from the log.
-: We'll not have the other details from MotionEvent
MotionEvent details cannot be recovered if not logged.
-: I can't find a way to get how long the kid hesitated (as @JackMostow mentioned in the comment)
When the trigger fires, you could log the time elapsed since the timer started. But to know how long the kid hesitated, we'd need to know when the kid finally stopped hesitating, i.e. did something. If all user actions get logged, we could compute the hiatus from when the timer started until the kid's first logged action thereafter, which we can recover after the fact from the log.
But we could do it at runtime by making every user action log:
the time elapsed since the previous user action and
the time elapsed since the previous tutor action,
which is probably a good idea anyway, if it's reasonably easy (i.e. doesn't require tracking down lots of different events to log).
(from discussion of logging hesitations): ... I'm glad it works as expected. I think I was looking at an older log where GESTURE triggered only once -- the log-only-once bug that you fixed. Here are my responses you requested: @JackMostow: I'm logging gesture from here. Or the previous method would suffice? Or both?! -- Without knowing the difference in what they log, I'd guess that either would suffice as long as the gesture gets logged. By doing it this way: +: We can have Current Tutor Id & Current Student Id Student id is constant throughout a session. Tutor id would be useful to have but can if necessary be inferred later from the log. -: We'll not have the other details from MotionEvent MotionEvent details cannot be recovered if not logged. -: I can't find a way to get how long the kid hesitated (as @JackMostow mentioned in the comment) When the trigger fires, you could log the time elapsed since the timer started. But to know how long the kid hesitated, we'd need to know when the kid finally stopped hesitating, i.e. did something. If all user actions get logged, we could compute the hiatus from when the timer started until the kid's first logged action thereafter, which we can recover after the fact from the log. But we could do it at runtime by making every user action log:
the time elapsed since the previous user action and
the time elapsed since the previous tutor action,
which is probably a good idea anyway, if it's reasonably easy (i.e. doesn't require tracking down lots of different events to log).
_Originally posted by @JackMostow in https://github.com/RoboTutorLLC/RoboTutor_2020/issues/42#issuecomment-856728276_