A Framework for Adding Packages to Graph Transformation Approaches

Graphs are a commonly used formalism for modeling many different kinds of static and dynamic data. In many applications, data modeling can be improved by using hierarchically structured graphs. But, while there already exist hierarchical graph data models, no general-purpose hierarchical graph data model exists yet, which unifies common features of these domain-specific models. In this paper, we present graph packages, a general formalism for defining hierarchical graphs, supporting the most important features found in known applications. Because of the dynamic nature of graphs, hierarchical graph transformation is also an important issue to be dealt with when using hierarchical graphs. Motivated by the successful application of graph grammars to the specification of graph transformations, we also introduce a framework that allows to specify hierarchical graph transformations by combining existing graph grammar approaches and our graph package concept. These concepts are a step towards the definition of a general-purpose hierarchical graph data model.