Programs that can recognize students' hand-drawn diagrams have the potential to revolutionize education by breaking down the barriers between diagram creation and simulation. Much recent work focuses on building robust recognition engines, but understanding how to support this new interaction paradigm from a user's perspective is an equally important and less well understood problem. We present a user study that investigates four critical sketch recognition user interface issues: how users integrate the process of triggering recognition into their work, when users prefer to indicate which portions of the diagram should be recognized, how users prefer to receive recognition feedback, and how users perceive recognition errors. We find that user preferences emphasize the importance of system reliability, the minimization of distractions, and the maximization of predictability.

No abstract