Latency (time from doing something physically to seeing response on display)? And how much is it limited by the design of the system and how much by processing power?

 

Follow up question if he wants to discuss it: Will going for lower latency lower the accuracy/usability?