Title: On Specifications and Proofs of Timed Circuits

Abstract: Given a discrete-state continuous-time reactive system, like a digital circuit, the classical approach is to first model it as a state transition system and then prove its properties. Our contribution advocates a different approach: to directly operate on the input-output behavior of such systems, without identifying states and their transitions in the first place. We discuss the benefits of this approach at hand of some examples, which demonstrate that it nicely integrates with concepts of self-stabilization and fault-tolerance. We also elaborate on some unexpected artefacts of module composition in our framework, and conclude with some open research questions.