The scheme followed until now by the team has always been the same. It has been illustrated by the work on model transformation for example. First we work on an innovative research idea to get it accepted by the community. Then we promote it towards partial normalization and we implement a proof of concept prototype. Following this, we publish an improved version of the prototype as an open-source contribution. Until now we have used the Eclipse foundation to this purpose. When this is established, the improved prototype has to find and gather a significant user community to show its adequacy to user needs. In case the approach is successful, it needs to be stopped at this point because a research team is not formatted to support a large user community. The solution experimented at this stage consists in making arrangements with industrial partners so that they may help in supporting the user community while the research resources of the team may be used to start an additional software prototype cycle.
The most ancient tool developed in the team is the ATL model transformation language. In this case the transfer state has been recently reached and we are presently working on the last phase of the process. The idea is to allow at the same time stabilization of the tool and ongoing research. Since these two objectives are in contradiction we have found a feasible solution. As the usability of the current version will be improved by an industrial partner in the ATL initiative, other new research branches will be opened in the ATL-Research initiative. Needless to say, both initiatives will continue to be based on open-source practices.
The other prototypes developed by the team have not yet reached the same recognition level, but they are following a similar path. We name AmmA (see 5.7 ) the set of software tools developed by the AtlanMod team.