Abstract
The economies of numerous nations, including New Zealand, rely heavily on the contributions of small and medium enterprises (SMEs). In the last decade a shift to the provision and use of software services, rather than products, has seen these offerings developed and delivered by multiple parties working in distributed and possibly transient networks, indicating that software SMEs could not only survive but can, in fact, thrive long-term at a small scale.
Most research on software success and failure, however, has been conducted in large organizations working on large-scale projects. Given the new work context described above, it is imperative that we also understand the circumstances that enable SMEs to be successful in software development. In general terms, software measurement has been considered to be one of the keys to success. The use of measurement is common in everyday life, such as checking the shortest path to your destination or weighing your luggage before taking a flight. It is also accepted in many science and engineering disciplines such as electrical and mechanical engineering. In spite of its acknowledged importance and rather obvious benefits, the use of measurement in software engineering has been variable. Software measurement is particularly conceived as a complex procedure, and therefore as a challenging and potentially costly endeavour, by software SMEs (SSMEs).
The aim of this research is to comprehensively understand and then contribute to more effective planning, deployment, operation and management of measurement programs specifically in the context of SMEs. The research reported in this thesis explores in detail the particular challenges that are encountered by SMEs when they embark on a software measurement initiative. It then proceeds to identify ways in which SMEs could effectively and efficiently implement light-weight software measurement programs (SMPs). In doing so the research combines elements of observation, design, intervention and evaluation under the umbrella of a Design Science Research (DSR) methodology. In three DSR the research conducts 1) problem identification through a literature review (via a mapping study) and an industrial review (via practitioners’ interviews), 2) solution design through field studies, and 3) evaluation through a survey. A mix of quantitative and qualitative methods is used as appropriate for each phase.
In the first phase, a comprehensive systematic mapping study is first conducted to review prior literature that had addressed SMPI in SMEs, to understand the state-of-the-art. The mapping study leads to the identification of a research gap which is further investigated through an industrial review. In the industrial review, 22 face-to-face interviews are conducted with professionals from SMEs. Hereafter, data analysis methods based on Grounded Theory (GT) enable the development of exploratory frameworks of four aspects of software measurement program implementation – challenges, obstacles, benefits and success factors – which forms one of the primary contributions of this research.
Based on these intermediate findings the second phase of this research involves the development of a novel framework which is intended to overcome (or at least reduce the severity of) measurement implementation challenges faced by SMEs. Implementing and sustaining a framework for the efficient planning and management of measurement programs remains a challenge for many software organizations, and particularly SMEs. Therefore, in this research phase, a comprehensive framework is proposed and refined, based on field studies that consider its adequacy in relation to the identified challenges and obstacles. The framework, referred to as the Software Measurement Framework for SMEs (SMF4SME), as developed and enhanced in the field over three cases, is a further novel contribution of this research.
The last major research phase validates the SMF4SME by seeking the insights of a sample of software practitioners working in SMEs, with respect to its perceived usefulness. An industrial survey is designed and distributed to potential participants to get their feedback. More than 100 respondents provide favourable indications regarding the coverage and potential utility of the framework in SSMEs.
Overall, this research work contributes to both theory and practice by providing an improved understanding of SMPI in SMEs along with a validated SMF4SME intended to overcome (or at least reduce the severity of) measurement implementation challenges in SMEs.