Article Preview
TopIntroduction
Software process improvement (SPI) is a rather old idea – probably as old as the field of software engineering. With the advent of the Capability Maturity Model (Humphrey, 1989; Paulk, Curtis, Chrissis, & Weber, 1993; Curtis, 2000) and a little later Software Process Improvement and Capability Determination (El Emam, Drouin, & Melo, 1998) the concept of SPI took a more formal form. Later again there has been an increased focus on the competences of software engineers and in software companies (Colomo-Palacios, Tovar-Caro, García-Crespo, & Gómez-Berbís, 2010; Aramo-Immonen, Bikfalvi, Mancebo, & Vanharanta, 2011). Despite this there have been remarkable few studies of the diffusion of SPI.
A review of literature searched in high-level outlets through Web of Knowledge and Google Scholar shows few but important research contributions. The Process Maturity Profile compiled by SEI (Software Engineering Institute, 2009) summarizes data from organizations voluntarily reporting their maturity levels. Reporting an appraised CMMI level does not necessarily imply an ongoing SPI practice, but it is a reasonable assumption that reporting organizations are interested in SPI. Since the report gives no information on the size of the population from which the reporting organizations could be seen as a sample it is not possible to deduce the percentage of software organizations practicing SPI. However, the report does give a minimal absolute number of organizations. Since the report has been produced regularly for seven years, it can describe trends. In all 3906 organizations have reported their appraisals by July 2009. 92% of these are on levels 2-5. Of the 3863 organizations reporting size data, 85% have more than 26 employees in the area of the organization that was appraised. As the sample increases there is a trend towards more organizations being on level 3 and fewer on 5. Only a couple of Danish organizations have reported appraisals to SEI, and only one of these is in our population.
The research question we are pursuing in this paper is: How many software companies in Denmark are engaged in software process improvement, what is the effect, and how is the effect related to the companies’ situation?
The paper is structured as follows. The theories we build on are presented together with our model of SPI diffusion. We present the research method with a particular focus on determining the population and a reasonable sample. We then analyze the empirical data and discuss the findings. Finally, we conclude the paper.