"Talking tests": a Preliminary Experimental Study on Fit User Acceptance Tests