THE SOFTWARE ARCHITECTURE PROCESS

The bulk of the work taking place within the realm of software architecture has been geared toward the definition of a set of common architectural styles (Garlan and Shaw, 1996) and the development of languages and techniques for expression and analysis of architectures belonging to each style (e.g., UniCon in Shaw et al., 1995), as well as toward developing special languages to facilitate domain-specific architectural description and analysis such as Rapide (Luckham et al., 1995a), MetaH and Control-H (Binns et al., 1995), SDFR (White,1995a,1995b), and DFR (White, 1996) to name a few. This paper explores not another architectural style, notation or analysis technique, but rather we present a generic view of the process of software âarchitectingâ, and the role software architecture plays within the software development organization. In particular we identify three tasks that we feel are the defining central activities of the software architecture process: architecture extraction, architecture generalization, and architecture reuse. We also show the existence of the above process tasks, though not always explicit, within new life cycle models that have emerged in order to facilitate reuse and reengineering. 1