Supporting reuse and Configuration : A port based SCM model

In this paper we present a Software Configuration Model (SCM) that has been specifically developed to consistently support reuse of software components and configuration of systems. In designing the model we explicitely wanted to address the following issues: - supporting modular software development and facilitating its modifiability; - supporting reusability of components and definition of reuse techniques; - supporting definition of sets of equivalent (with respect to a certain property) components (variant/version control); - supporting configuration of stable parts of a system. The approach we have taken follows the experience we gained in supporting configuration management on a (logic-deductive) data base. That is, we define a schema for a kernel configuration environment data base, (i.e. the objects, their relationships and the static constraints), and a set of operations which represent the dynamic component of the data model and among which configuration tools are defined. The schema we propose is very general and the design focussed on the definition of a kernel data structure that maintain the information, relevant to configuration activities, produced during the various phases of the software life cycle; basic notions of our schema are those of module and variant group. A module is an elaboration unit characterized by an interface which describes the functionalities it needs in order to make available to the external world those ones it produces. Thus, an interface completely characterize a module by abstracting with respect to its internal topology. In this way, in the external context, a module is not different from an atomic structure.