I think the main difference is they are easy to interrupt and continue from the current real position, where doing that with CA is more difficult. That would let you do some nifty stuff with gestures, I guess.