Are Statecharts Finite Automata?

Statecharts, which have been introduced by D. Harel in 1987, provide compact and expressive visual formalisms for reactive systems. They have been widely used as a modeling tool and adopted by Unified Modeling Language (UML) as an important technique to model the dynamic behaviour of objects. One of the fundamental questions concerning statecharts is what the computation power of statecharts is. Until now, most descriptions consider that the computing power of statecharts is the same as that of Finite State Machines or Finite Automata, though no accurate arguments or proofs have been provided. In this paper, we show for the first time that the computation power of statecharts is far beyond that of finite automata. We compare statecharts with Interaction Machines introduced by P. Wegner more than ten years ago. We show that the Interaction Machines are the most accurate theoretical models for statecharts.