Detection of bimanual gestures everywhere: why it matters, what we need and what is missing