The goal of this work package is to consolidate economical and technical issues regarding the adoption or transition of system family engineering practices.

The objective of the work package is to merge the existing CMMI and CAFÉ process frameworks into a single system family maturity framework with an ample consensus from industry.

The current de- facto standard process maturity framework CMMI (earlier CMM) does not take system family development into account. The SEI publication Software Process Improvement and Product Line Practice: CMMI and the Framework for Software Product Line Practice, by L.G. Jones and A.L. Soule (CMU/SEI-2002-TN-012) does not indicate a substantial improvement from the SEI itself, which is both involved in CMMI and Product Line Practices.

On the other hand, the CAFÉ project and its predecessor ESAPS have addressed several engi-neering practices regarding system family engineering in depth, ranging from requirements to architecture, derivation and testing in the context of a specific domain.

We believe that only companies mature in system family engineering will succeed here.

The objective is to provide a set of techniques and methods that guarantee the quality of system families in an organisational context and addressing new challenges. Quality aspects have already been addressed in ESAPS and CAFÉ. For example much effort has been devoted to architectural assessment and family testing. The obtained results need to be integrated in the process context of the organisation.

At the same time, the technological and market evolution poses new problems to family development organisations (e.g. dynamic reconfiguration of systems by end-users, security issues, dependability from 3 rd party software. Service-based systems)

This Work package builds upon the MDA initiative for consolidating the system family engineering approach into a Model-Driven system Family Engineering approach (MDFE). Model Driven Family Engineering will provide, organise, and manage not only a repository of common assets, but also an explicit repository o f models including development know-how. A corollary to this is that stable standards on know-how can emerge and be shared as well. Model driven family engineering will provide full support for weaving, assembling, transforming and refining all the choices up to implementation models specific to the particular execution infrastructure.

The high-level models capture the core knowledge of the family, whereas the low-level models realise the requirements on given platforms, according decisions done during the derivation. ESAPS and CAFÉ activities on variability modelling, decisions modelling or product derivation were an initial attempt to introduce model driven engineering as a way to address system families aspects.

Support the merging of existing assets in to a single family. The business interest is to effectively use the software that is already available, and introduce software reuse on a very wide scale. Not only software, other assets are as important. We have to focus on systems as a whole. In many cases, several families already exist and in order to keep their maintenance and construction manageable and to reduce cost we have to introduce reuse over family boundaries. This means we have to group our families into populations. This is a next step towards a higher level of reuse. Because of business reasons each family stays in existence, and systems have to be produced within these families.

One of the most valuable assets extracted from the ESAPS and CAFÉ projects has been the building of the European system family community. In FAMILIES we want to continue this trend by giving the community the necessary tools to be aware of all the European, and world-wide initiatives in this topic. On the other hand, the results of FAMILIES have to be well known by tool providers for them to be aware of the real needs of the European industry to approach system family development.