The problem starts with capture. Since the objective of capture is keeping your head clear, the process results in my inbox containing two types of tasks—those I will do, and those I might do. - DaFacto
In my situation I've got a lot of location-aware tasks that can only be completed at a certain store or off-site location. This means I've got to implement a list of context per location so I can easily shift to the tasks at hand for the location I'm currently working.
But aside from these locations, my contexts are all over the place. So maybe implementing a simplified approach for all tasks that are not related to those specific locations could be a great approach.
Something to think about.