If-Next-Also-Else: Controlling Graph Transformation

Abstract: Graph transformation is a declarative, rule-based formalism. However, when using it in practice, it is convenient to schedule rules explicitly, and not just rely on their own application conditions, even though such scheduling adds an imperative flavour. This can be catered for by adding a layer of *control* on top of the rules.

In this presentation, I will demonstrate the power of an expressive rule control language, then discuss its semantics. I will argue that the usual, process algebraic concept of operational semantics does not adequately capture the notions of “try-else” and “as-long-as-possible” that are very popular in the context of controlled graph transformation; and I will propose an extension that replaces CCS-like symmetric choice into a more primitive, asymmetric version.