Traditional approaches to Software Process Improvement focus on measurement
and evolutionary change within the established Software Process. This appr
oach is attractive because it has fairly low risk but has high measurement
cost, and limited potential benefit because of the concentration on evoluti
on. At Lucas Varity we are piloting an approach to Process Improvement that
focuses on deploying innovative research directed to improving productivit
y and maintaining or improving the existing high product quality. This appr
oach appears to offer significantly greater gains in productivity but requi
res a new approach to Process Improvement. The main elements are:
A commitment to long-term strategic research to identify product related in
novations within the company to acquire new technology and members of staff
able to pioneer innovations.
A commitment to medium-term research in the form of carefully controlled Pr
ocess Improvement Experiments that are oriented to providing quantitative e
valuation of proposed innovations.
The development of staff capability in Design Process Reengineering to prov
ide the basis for Process Improvement experimentation.
We provide an overview of our approach to Design Process Reengineering and
illustrate it by describing an ongoing experiment in the deployment of form
al methods tools developed to support hardware design used in the design of
safety-critical software. Preliminary results indicate that the use of suc
h methods offer considerable advantages over conventional approaches. (C) 1
999 Published by Elsevier Science B.V.