Abstract
Implementing a model transformation is a very complex task and in an MDA process, chains of model transformations are usually built. When writing such a transformation chain, developers often need to have information on the previously applied transformations. Thus, disposing of a traceability framework enabling to gather information on the transformation behavior is an important feature for a transformation language. In this paper, we propose to implement a traceability framework in the Kermeta language based on a language independent trace metamodel.