Testing software modelling tools using data mutation

Abstract

Modelling tools play a crucial role in model-driven software development methods. A particular difficulty in testing such software systems is the generation of adequate test cases because the test data are structurally complicated. This paper proposes an approach called data mutation to generating a large number of test data from a few seed test cases. It is inspired in mutation testing methods, but differs from them in the way that mutation operators are defined and used. In our approach, mutation operators transform the input data rather than the program under test or the specification of the software. It is not a test adequacy measurement. Instead, it generates test cases. The paper also reports a case study with the method on testing a modelling tool and illustrates the applicability of the proposed method. Experiment data clearly demonstrated that the method can achieve a high test adequacy. It has a high fault detecting ability and good cost effectiveness.