What're the impacts of multi-touch technology for input method frameworks?

1. There maybe multiple focuses on a single desktop session. So, only the "focus in" or "focus out" maybe insufficient. We need a new primitive, "focus_change (from, to)" to track the focus changing. It's also an impact for windowing system.

2. There maybe multiple active input contexts. The assumption "get active input context" is not valid anymore. Every operation should be bound to a specified IC (or entire desktop). Some GUI helper (or Auxiliary Window, such as virtual keyboard) need be able to create multiple instances for requiring ICs.

Do you have any more thoughts of the impact?

Once you have no concept about multi-touch, have a look about MPX, and the demo vidoes.