Long-lived, architecture-based software systems are increasingly important. Effective process support for these systems depends upon recognising their compositional nature and the active role of their architecture in guiding evolutionary development. Current process approaches have difficulty with run-time architecture changes that are not known a priori, and dealing with extant data during system evolution. This paper describes an approach that deals with these issues. It is based on a process-aware architecture description language (ADL), with explicit compose and decompose constructs, and with a hyper-code representation for dealing with extant data and code. An example is given to illustrate the ease-of-use benefits of this approach.
R. Mark Greenwood, Dharini Balasubramaniam, Sorana