Test case generation for information systems using reverse engineering techniques

Testing has a very important role in Software development, implying about a half of the total cost. Its automation tries to minimize these costs and increase the productivity. Whilst there are many well-known solutions for some tasks (such as test case execution), test case and test data generation in information systems which use databases is particularly complex. In this article we propose the automatic generation of test cases to verify information systems that store their data in relational databases. For this we work on a metamodel based on UML (automatically extracted with reverse engineering mecanisms) that represents the data model and the business rules. Applying pattern-matching techniques on this metamodel we detect interesting test situation, for which we generate test cases and their corresponding test data in order to reach the defined coverage criteria on the data model.