I just finished the tutorial for geofencing on Android (http://developer.android.com/training/location/geofencing.html) and I wonder why the 'callback' for geofences are done via pending intents and not a simple callback interface.
If implemented in an activity, one would usually disconnect the location client in onPause()
anyway, so previously added geofences would not be tracked either after the application paused/was destroyed, so why a pending intent? Or am I mistaken here?
This answer can be outdated - accuracy and realiability of google play services has changed a lot from it's initial release.
Some of my experiences with geofencing below. First of all - the main advantage of this technology is VERY low battery usage. In the fact, I can't notice any changes in battery life. It's really impressive. Service seems to use only Wi-Fi and network location. I didn't notice GPS running at all. I can't say if it's only hidden location icon or really not using GPS. Accuracy - it's terrible. 20 circle areas are not detected at all, except range of my home ap. It looks like whole position circle, including error must be inside of fenced area. 1000m areas are detected sometimes and with huge latency. Those experiments where made in open area with very low number of Wi-Fi ap around. I'm still trying to find really reliable settings foot this service. After getting intents I want to turn on GPS location and make final approach in my own code.
Mostly because geofences are designed to work even without your application running.
I believe that you are mistaken here. In fact, geofences specifically are not designed for directly triggering UI, as is discussed in the documentation:
Now, you might elect to say that you want to only use geofences while you have your activity in the foreground. However, you would have to remove those geofences in
onPause()
. A geofence will remain registered until its expiration time or manually removed, AFAICT.