Software Evolution

SE is an action in Valoria lab.

Members

Old Members

Research topics

    An intrinsic characteristic of software, addressing a real world activity, is the need for evolving to satisfy new requirements. The first law of Lehman, resulting from empirical studies, stipulates that a software must necessarily evolve else it becomes gradually less satisfactory in use. Maintenance activities represent, without surprise, a very significant proportion in companies developping software sales (from 50 to 75%, according to studies). Thus, it is necessary to propose methods, techniques and tools facilitating these activities while reducing their costs. This research group proposes to meet this need in the field of object-oriented and component-based software.

    We carried out some works to decrease the coupling between client and server objects through the concept of Service Group.  This increases, consequently, servers maintainability (Service Group project). Other work has also allowed to propose tools and methodology for evolving dynamically server objects using an instance inheritance mechanism (Adapter project). Currently, we are working on the proposition of a formal framework to guarantee the permanence of nonfunctional caracteristics (particullary, quality attributes) during the evolution of component-based applications (Cell project). This study, in long term, should be able to offer means of defining and checking a strong typing in components world beyond the syntactic aspects (as in typed languages, like Java) and functional aspects (as in Eiffel Language with pre/post-conditions and invariants) because taking into account nonfunctional aspects.

Former projects

• Service Group: A programming paradigme for distributed objects,
  

• Adapter: Dynamic adaptation of distributed objects,
  

Projects in progress

• CELL: Contracting Software Evolution,
  

• AURES: Software Evolution Assistant,
  

• Substitute: Software Component Selection and Substitution,