After briefly reviewing the difficulties of achieving major improvement in
the process of global software development and maintenance, software evolut
ion, the paper introduces the FEAST hypothesis. This states that such probl
ems may, at least in part, be due to the feedback nature of that process an
d that this is, generally, overlooked in seeking such improvement. Results
to date from the FEAST/1 project support the hypothesis and also the laws o
f software evolution as recognised and refined over some 25 years. The proj
ect is exploring the phenomenon in depth by identifying, modelling and simu
lating the evolutionary behaviour of a number of industrial software projec
ts using both black box and system dynamics techniques. It is now demonstra
ting the presence and impact of feedback on process behaviour and improveme
nt, and deriving guidelines for software process planning and management. (
C) 1999 Elsevier Science Inc. All rights reserved.