Résumé
In this extended abstract, we report an ongoing experience conducted during a Master project on the migration of two industrial software systems. The project was proposed by a major IT service company (not cited here for confidentiality reasons) which would like to investigate a migration processing chain, in order to renovate legacy software composed of man-machine interfaces, databases and procedural source code. The aim of the renovation is to migrate to the object-oriented paradigm and generate new source code. Due to the limited time that the Master students had in their curriculum for this project, we restricted the study to the extraction of a class model. A processing chain was proposed and a few heuristics, taken in the literature, have been tested, with no really conclusive results. We report here the current status of the project in order to get feedback and new ideas to build for the future. Software renovation still remains a costly and time-consuming process for IT service companies, that can be viewed as a waste of resource, compared to the development of new software. Nevertheless, if timely and effective measures are not taken to regularly update the design, the source code, the documentation and all related artifacts, it may arrive the day where the software can no longer be understood by human, or compiled by the new compilers, or even ran on the new servers. The challenge is to maintain and migrate with a low cost the legacy software, before real problems arise. In this extended abstract, we report an The authors would like to thank the IT service companies that brought the renovation project and followed the work in progress and the master students (Luc Debène, Chaymae Regragui and Cedric Cambon) that made a tutorial for the use of Famix in the context of this project.