Taipale and Smolander report on an extensive qualitative study of software testing. They conducted 41 interviews with development and testing managers, testers, and system analysts. In total, 946 pages of transcribed interview data were analyzed, using the grounded theory approach and ATLAS.ti software. The analysis yielded seven dimensions that were used to distinguish five representative organizational units (called cases). For example, the dimension “level of business orientation” varied from purely service oriented (case E), to purely product oriented (case D).
The authors discuss various cause and effect pairs. For example, greater use of software components creates a demand for components that are more testable. The authors also list five process improvement propositions. Proposition 1 states that testing should be adjusted according to business orientation. Proposition 2 calls for enhanced testability of software components. Proposition 3 calls for efficient communication and interaction between development and testing. Proposition 4 calls for the early involvement of testing and planning for testing. Proposition 5 calls for the use of risk-based testing.
The authors ambitiously set out to establish a theory of software testing from the ground up. Disappointingly, the emergent five propositions are largely already well understood. In addition, the authors should have critically reflected on the qualitative methodology they used, since analyzing 946 pages of transcripts is far from light work. To establish a theory of software testing, building a business simulation model based on the existing extensive literature on software testing might have been more productive. This paper will be of passing interest, only to those involved in software process improvement.