Improving software process modelling and enactment techniques

Abstract

Most of the work previously carried out on software processes addressed the problem of modelling as a monolithic description focusing on one perspective e.g. product, process, role, etc. Little has been done to address the problem of software process modelling through the dynamic composition of consistent and reusable pre-composed components. To build software processes from components, there are many concerns to take into consideration:

• 1) building a software information base of reusable components,

• 2) determining the components that make up a process model, while this composition is achieved dynamically to assure the mutual consistency of components,

• 3) how to achieve integration to support full product, activity and role perspectives,

• 4) how to incrementally construct the model by introducing new components or updating component versions in order to deal with evolution.

Our aim is to provide answers to these questions by proposing a new model that is designed as part of the Pynode project. The proposed model integrates these aspects in a unified manner. This work is the result of our five years' experience on our systems: Tempo/Adele and Opsis. This paper presents the main components of the model, describes their structure and explains their role in supporting software processes.