A Smart Structured Test Automation Language (SSTAL)

In model-based testing, abstract tests are designed in terms of a model, for example, a path in a graph and concrete tests are expressed in terms of the implementation of the model. Two problems exist while converting abstract tests to concrete tests: the "mapping\´\´ problem and test oracle problem. Abstract tests cannot be applied directly to the actual program and they must first be mapped to concrete tests. Testers currently solve this mapping problem by hand. If one basic action is used multiple times in different abstract tests, testers must write a lot of redundant code for the same action. This process is time-consuming, labor-intensive, and error-prone. For the "mapping\´\´ problem, this research will design a structured test automation language to partially automate the mapping from abstract tests to concrete tests. First, programmers or testers use the automation language to create mappings from each basic identifiable part (e.g. an action in a finite-state machine) of the model to the corresponding executable programming code. Once the mappings are generated, concrete tests can be generated automatically from the abstract tests. This structured test automation language will be used to improve the efficiency of generating concrete tests from a variety of abstract tests and reduce the potential errors. Test oracle problem refers to how to determine whether a test has passed or fail. While writing executable code, it is difficult to decide which parts of the program state should be compared by the concrete tests. A guideline will be established to evaluate stats of the program effectively and efficiently.