Mobile Application Development: Automated Test Input Generation Via Model Inference based on User Story and Acceptance Criteria

Abstract

In the past few years, there has been observed explosive growth in the development of Mobile Applications across Android and iOS operating system which has led to the direct impact towards mobile app development. In order to design and propose quality-oriented apps, it is the primary responsibility of the developers to devote time and sufficient efforts towards testing to make the Apps bug free and operational in the hands of end users without any hiccup. In order to test the mobile apps, manual testing procedures takes prolonged amount of time in writing test cases and even the full testing requirements are not met. In addition to this, lack of sufficient knowledge by the tester also impacts overall quality and assurance that app is bug free. To overcome all the issues of testing, and to assure that apps designed by developers are almost bug free, we propose a new testing methodology cum tool “AgileUATM” which works primarily towards white-box and black-box testing. With this tool, all the test cases are generated automatically based on user stories and acceptance criteria by using formal specification and Z3 SMT solvers. To test the validity of the proposed tool, we applied the tool in real-time operational environment with regard to test Mobile apps. Using this tool, all the acceptance criteria is determined via user stories. The testers/developers specify requirements with formal specifications based on programs properties, predicates, invariants, and constraints. From the results, it is observed that the proposed tool i.e. AgileUATM generated effective and accurate test cases, test input, and expected output was generated in a unified fashion from the user stories to meet acceptance criteria. In addition to this, the tool also reduced the development time to identify test data as compared to manual Behavior Driven Development (BDD) methodologies. With this tool, the developers got better idea with regard to required tests and able to translate the customers natural languages to the computer language as well.