Towards MDA Software Evolution

Abstract

This chapter discusses software evolution, challenges and strategic directions in the context of MDA. Various authors agreed that it is difficult to define completely software and then, software evolution. Software is certainly more that bits stored in a file, it is an abstract idea that encompasses the concepts, algorithms embodied in the implementation as well as all its associated artifacts and processes. Research seems to confirm that computer software and process software have much in common. Osterweil (2003) assures that software processes are software too. In other paper (Osterweil, 2007), he suggests analyzing the nature of software and proposes to define taxonomies for exploring characteristics and approaches to the development, verification of qualities and software evolution. The exploration of these questions is an important current of software engineering research. On the other hand, evolution is defined as a process of gradual change and development from fewer and simpler forms to higher, more complex, or better ones. In biology, evolution is related to develop over time often many generations, into forms that are better adapted to survive changes in their environment. Thus, evolution captures the notion of something improving and changes occur in species in successive generations, i.e. individuals get old and species evolve. Jazayeri (2005) analyzes the definition of software evolution. The concept of “specie” in software may be associated to meta-levels describing families (species) of software systems. These meta-levels or architectures are created as improvements to previous existing ones and describe evolved families of software systems.