Improving research methodology as a part of doing software process improvement

Working with software process improvement (SPI), using measurements and statistical methods, can easily be a challenge for SPI people. Workers dislike having their processes monitored and managers dislike being informed of their processes being less than perfect. The power needed by SPI people for carrying out their work may thus be challenged, and one way of facing up to this might be to argue with workers and managers through authority established by SPI literature, SPI practice networks and SPI research networks. There are similarities between the specify-produce-control cycles of software development and the hypothesize-experiment-test of scientific research. If SPI people choose to ally with academia by becoming part of academia, how can they make trivial SPI problems become interesting academic problems? How may the abstract and elaborate process of producing new academic knowledge be of practical use for downto-earth SPI? Based on Humphrey’s initial design ideas for the capability maturity model (CMM), and using the current ISO 9004:2000 as a framework for self-control and – improvement for the SPI people, this paper suggests a simple way for aligning industry and academia through a way of doing action research specifically designed to fit with the CMMand ISO-standards.