Utilizing domain specific modelling for software testing

Automated execution of manually defined regression tests is a very widely used and well-known area. While test execution can be more easily automated, test case creation and maintenance are still mainly manual efforts and practically the biggest cost factors in software testing. We view writing test cases as basically a programming activity and believe it can thus benefit from extended application of generic programming tools and techniques. In this paper, we describe our work in applying domain-specific modelling (DSM) to the domain of test case creation. DSM is a variability handling method typically applied in software development. It is widely used and powerful method best applied when there are several kinds of variations. DSM is typically tailored to make own optimized modelling solution inside a company, after which it can be applied effectively and without requiring specific programming skills. In this paper we describe how we have applied DSM to describe variability in software behaviour in terms of test cases, and its application in a case study. The results show a reduction in the cost of over test automation.