Mining Data for Ecore Libraries

In model-driven development software systems consist not only of code (and maybe documentation) as central artifacts, but also of models of the software system that is to be built. These models are frequently created in UML and oftentimes enable automatic code generation. However, in contrast to the relatively common practice of reusing code, models created during the development of a software system are usually not reused. The Model Composition and Combination Vault (MoCCa) provides a system to link and store models in a way that enables easy reuse of models to facilitate and speed up the process of modelling new software systems. One of the problems when trying to reuse models is the need for an as-big-as-possible set of models that are already known. This set of known models ? MoCCa uses a database for this ? has to be created by hand so far, making the process tedious and time-consuming. The objective of this thesis is to create a way to semi-automatically refine available models and store them in the database of known models. In order to do this the system should identify parts of the model that are already known and separate those from the unknown parts. The unknown parts should then be decomposed into meaningful units which are finally stored in the database. Such a system would then enable a rapid inclusion of old models into the system to reuse them in new projects.

Project information



Thesis for degree:



Nils Sewing