Adaptation of Models to Evolving Metamodels
Abstract
The problem of automatic or semi-automatic adaptation of models to their evolving metamodels is gaining importance in the Model-Driven community. Recent approaches propose to adapt models using predefined information (i.e., a trace of changes). Unfortunately, this information is not always available in practice. In many situations metamodels evolve without keeping track of the applied changes. We propose a more general two step solution. First step computes equivalences and differences between the metamodels and saves these into a ``weaving model''. This weaving model acts as a high-level specification of adaptation transformation. Second step translates this model into an executable transformation. This technical report shows the results obtained in applying the approach on two concrete scenarios: a Petri net metamodel, and the Netbeans Java metamodel.
Origin : Files produced by the author(s)
Loading...