On the Power of Play-Out for Scenario-Based Programs

Abstract

Abstract. We investigate the power of play-out, the execution mechanism associated with scenario-based programming, which was defined as the operational semantics of live sequence charts (LSC). We compare some of the play-out strategies and mechanisms suggested in the literature, and discuss their strengths and limitations. Specifically, we define a simple infinite hierarchy of LSC programs, and use it to show that smart play-out, the lookahead version of play-out guided by model-checking, is strictly weaker than full synthesis from LSC.This paper is dedicated to Prof. Willem de Roever, friend and colleague, with admiration and respect for his scientific achievements, his impact on the community, and his boundless energy.

Compositional Synthesis of Reactive Systems from Live Sequence Chart Specifications

“…Another example is adopted from [12], where it is shown that synthesis is strictly stronger than smart play-out. We modify the example to form a series of inconsistent specifications of growing lengths, where specification i requires considering i supersteps ahead in order to prove its inconsistency.…”

“…However, looking only one superstep, or a finite number of supersteps, ahead, is not sufficient either. An example for this is given in [12]. This leads to the synthesis problem, i.e., given an LSC specification, finding a reactive system that adheres to the specification, or proving one does not exist.…”

Controller Synthesis from LSC Requirements

“…This approach uses verification methods-in particular, model-checking-to run LSC specifications and avoid certain violations that may occur in the original version of play-out. Unfortunately, smart play-out cannot avoid all possible violations [7]. This paper addresses an improvement to smart play-out which guarantees non-violation over all computations, provided that the requirements are realizable.…”