Emerald Group Publishing Limited
Copyright © 2000, MCB UP Limited
Keywords: Automation, Cybernetics, Research, Technological developments
In the UK two projects have been supported to help us better understand and consequently, it is hoped, improve the way in which software is developed and maintained. This is the project, leaders say, to help accommodate "the agile and highly adaptive enterprises of the new millennium that are already demanding more and more flexibility from their software support". This is further expanded in the project aims which say that:
The potential for organisations to improve their processes, and over-all effectiveness, is very much related to their increasing reliance on computers for all aspects of their activity. Such highly critical technology must become not only more reliable but also more adaptable. After all, business effectiveness and maintenance of a company's competitive edge strongly depends on adequate and timely evolution of individual software systems so that they remain satisfactory under changing operational and environmental conditions. Computer systems and the business activities they support comprise an integrated system. As the number of applications grows, the system components become ever more intertwined and interdependent. A change to any one can have an unexpected impact on almost any other aspect of business operation and effectiveness. Recent years have therefore seen increasing worldwide interest in disciplined approaches to business process analysis so that companies' systems become flexible enough to support their business operations in an ever-changing operational environment.
Professor Lehman and his team called their project - feedback, evolution and software technology - FEAST, and their hypothesis was explored in their first research council supported project, FEAST-1. They say that it:
confirmed that software systems evolve in a disciplined fashion, but to understand this process requires one, for example, to understand how the feedback of a "need to change" is managed. The presence and influence of feedback, and its impact on the software process, was first noted during a study of the IBM OS/360 operating system in the 1970s when it was also shown that continuing evolution is an essential part of real world computer applications and software. This observed behaviour, supposed by a developing theory, became encapsulated in a series of software evolution laws.
Professor Lehman formulated his FEAST hypothesis projects in 1994 with the FEAST projects following. They were:
FEAST/1 studied data from a number of evolving systems provided by collaborators: ICL Logica and Matra-BAe Dynamics. Through Dr D E Perry in his capacity as a Senior Visiting Fellow to the project, other data were obtained from Lucent Technologies. This examination of the role of feedback in the software process begins a long-term and difficult investigation - being continued in FEAST/2 - into how software systems can reflect the demands modern business places on them. The results are producing technology and tools for planning and controlling software evolution, and a theoretical framework for better mastery of software evolution and the software process. FEAST/2 will test the theory of this approach against a wide range of processes and applications and determine, in particular, which feedback mechanisms are effective in controlling the many inevitable complex interactions. For a world becoming ever more dependent on computers, and therefore on software, study of such issues represents a vital challenge. If upheld in the face of advancing software technology, exploitation of the 1970s observations should have profound impact on the approach to and implementation of software and other business processes.
Both projects were funded by the UK's Engineering and Physical Sciences Research Council (EPSRC). Further information is available on the Web site: www-dse.doc.ic.ac.uk/-mml/feast/