In modern model based development processes models are part of the operative
software. Hence, like with source code, many versions of models arise. Therefore
one requires powerful tools for comparison and mixing of models. Such tools are based
on algorithms which calculate the differences between models. High demands are set onto
such difference algorithms: They must regard the structure of the specified model type and
the features of typical edit processes whilst being preferably fast and also depict the
differences between the given models in a clever way. - In other words, they should deliver
differences of high quality. However, these different pretensions are contradictory
to each other: An optimization for quality impairs the – often already high – runtimes further
on. Because of this reason an acceptable compromise between quality and effort must be found for
the respective task. Such compromises cannot be discussed here because systematic approaches for
quality definition of differences are missing. Major goal of the QuDiMo-Project is the
definition of exactly such quality measures for model differences as well as an approach to
depict the optimization of existing difference algorithms concerning both runtimes and the
achieved quality of differences.
A detailed overview of the pursued objectives can be found here.