Towards a Framework for Information System Testing: A Model-driven Testing Approach

摘要

Testing has an important role within the development of a software system; the automation of testing tasks has been largely used with the goal of minimizing costs and increasing productivity. For some of those tasks-as it is the execution of test cases-well-known solutions already exist as the industry adopted them many years ago. This is not the case with test data generation, and even less for software that uses databases, where this task is particularly complex. In the present work we propose to generate test cases automatically to test information systems that use relational databases. We represent the data model with UML Data Modeling Profile, automatically extracted from the database with reverse engineering techniques. Then, applying model-driven testing, test cases are generated from the data model, represented with the standard UML Testing Profile. The test case generation also includes test data inputs, in order to accomplish certain coverage criteria defined on the schema.