Model transformation languages relying on models as ADTs

Abstract

In this paper we describe a simple formal approach that can be used to support the definition and implementation of model to model transformations. The approach is based on the idea that models as well as metamodels should be regarded as abstract data types (ADTs), that is to say, as abstract structures equipped with a set of operations. On top of these ADTs we define a minimal, imperative model transformation language with strong formal semantics. This proposal can be used in two different ways, on one hand it enables simple transformations to be implemented simply by writing them in any ordinary programming language enriched with the ADTs. And on the other hand, it provides a practical way to formally define the semantics of more complex model transformation languages. © 2010 Springer-Verlag.