Towards a Test Automation Improvement Model (TAIM)

In agile software development, industries are becoming more dependent on automated test suites. Thus, the test code quality is an important factor for the overall system quality and maintainability. We propose a Test Automation Improvement Model (TAIM) defining ten key areas and one general area. Each area should be based on measurements, to fill the gap of existing assessments models. The main contribution of this paper is to provide the outline of TAIM and present our intermediate results and some initial metrics to support our model. Our initial target has been the key area targeting implementation and structure of test code. We have used common static measurements to compare the test code and the source code of a unit test automation suite being part of a large complex telecom subsystem. Our intermediate results show that it is possible to outline such an improvement model and our metrics approach seems promising. However, to get a generic useful model to aid test automation evolution and provide for comparable measurements, many problems still remain to be solved. TAIM can as such be viewed as a framework to guide the research on metrics for test automation artifacts.