I wonder if we could build better implicit state machines if we could more easily visualize our imperative code by annotating it with visualization hints and parsing the AST. Also easier to debug/step-through this way.