Software testing is an essential knowledge area required by industry for software engineers. However, software engineering students often consider testing less appealing than designing or coding. Consequently, it is difficult to engage students to create effective tests. To encourage students, we explored the use of gamification and investigated whether this technique can help to improve the engagement and performance of software testing students. We conducted a controlled experiment to compare the engagement and performance of two groups of students that took an undergraduate software testing course in different academic years. The experimental group is formed by 135 students from the gamified course whereas the control group is formed by 100 students from the non-gamified course. The data collected were statistically analyzed to answer the research questions of this study. The results show that the students that participated in the gamification experience were more engaged and achieved a better performance. As an additional finding, the analysis of the results reveals that a key aspect to succeed is the gamification experience design. It is important to distribute the motivating stimulus provided by the gamification throughout the whole experience to engage students until the end. Given these results, we plan to readjust the gamification experience design to increase student engagement in the last stage of the experience, as well as to conduct a longitudinal study to evaluate the effects of gamification.

Entity reconciliation (ER) aims to combine data from different sources for a unified vision. The management of large volumes of data has given rise to significant challenges to the ER problem due to facts such as data becoming more unstructured, unclean, and incomplete or the existence of many datasets that store information about the same topic. Testing the applications that implement the ER problem is crucial to ensure both the correctness of the reconciliation process and the quality of the reconciled data. This paper presents an approach based on model-driven engineering that allows the creation of test models for the early integration testing of ER applications, contributing in three main aspects: the description of the elements of the proposed framework, the definition of the testing model, and the validation of the proposal through two real-world case studies. This validation verifies that the early integration testing of the ER application is capable of detecting a series of deficiencies, which a priori are not known and that will help to improve the final result that the ER application offers.

Functional testing of applications that process the information stored in databases often requires a careful design of the test database. The larger the test database, the more difficult it is to develop and maintain tests as well as to load and reset the test data. This paper presents an approach to reduce a database with respect to a set of SQL queries and a coverage criterion. The reduction procedures search the rows in the initial database that contribute to the coverage in order to find a representative subset that satisfies the same coverage as the initial database. The approach is automated and efficiently executed against large databases and complex queries. The evaluation is carried out over two real life applications and a well-known database benchmark. The results show a very large degree of reduction as well as scalability in relation to the size of the initial database and the time needed to perform the reduction.

Testing a database application is a challenging process where both the database and the user interaction have to be considered in the design of test cases. This paper describes a specification-based approach to guide the design of test inputs (both the test database and the user inputs) for a database application and to automatically evaluate the test adequacy. First, the system specification of the application is modelled: (1) the structure of the database and the user interface are represented in a single model, called Integrated Data Model (IDM), (2) the functional requirements are expressed as a set of business rules, written in terms of the IDM. Then, a MCDC-based criterion is applied over the business rules to automatically derive the situations of interest to be tested (test requirements), which guide the design of the test inputs. Finally, the adequacy of these test inputs is automatically evaluated to determine whether the test requirements are covered. The approach has been applied to the TPC-C benchmark. The results show that it allows designing test cases that are able to detect interesting faults which were located in the procedural code of the implementation.

Data mining studies need a set of algorithms to analyze the data and a way to administer the data under study. In the industrial environment this data (specially low level data) is usually stored at several local databases spread over the factory installations. This paper describes a novel attempt to set up a data mining engine in a steel factory installation based on new information technologies (Internet network, application servers, etc.) to facilitate data mining studies involving such scattered data. Its main benefits and drawbacks are addressed.

The use of techniques for automating the generation of software test cases is very important as it can reduce the time and cost of this process. The latest methods for automatic generation of tests use metaheuristic search techniques, i.e. genetic algorithms and simulated annealing. There is a great deal of research into the use of genetic algorithms to obtain a specific coverage in software testing but there is none using the metaheuristic Tabu search technique. In this paper, we explain how we have created an efficient testing technique that combines Tabu search with Korel chaining approach. Our technique automatically generates test data in order to obtain branch coverage in software testing.