To advance the state of the art of software engineering, our research group is active in the following fields of research:

Test Languages

A strong focus of our research group is the analysis and improvement of software quality. One pillar of our research in this area is the work on test languages. Prof. Grabowski contributed to the development of the two test languages TTCN-3 and the UML Testing Profile (UTP). We are still active in the maintenance of TTCN-3 and the development of tools for the maintenance of TTCN-3 test suites. Furthermore, we are heavily involved in the development of the Test Description Language (TDL).

Language Evolution

Programming languages need to evolve in order to keep up with changing technological circumstances and developer needs, otherwise they may face a decrease in their usage and popularity. New language features and paradigms aim to enable more secure, more reliable, and more efficient software and software development. However, it is difficult to estimate how new language features will be perceived by developers, whether they will be adopted in practice, and how to encourage their adoption.

Software Engineering and Data Science

Empirical studies are an important aspect of software engineering research. In our research group, we apply data science methods to software engineering problems to improve and empirically assess different parts of the software development life cyle. This includes the development of novel software mining techniques, e.g., for defect prediction, developer-specific models, and simulations of software projects to create decision support systems for project managers. Furthermore, our interest also goes in the other direction: how can we use our software engineering skills to advance data science? We work together with domain researchers to develop data science software that answers their research questions, investigate the replicability of studies, and are interested the quality assurance of data science tools.

Mutation Testing

How do we assess and improve the quality of tests? This question can be addressed by applying mutation testing: artificial faults are seeded into the program, and any non-detected fault among these is an indication of the hole in the test suite. Test suite evaluation, prioritization, minimization, and generation can be achieved through mutation testing. Our focus is related to the different challenges and applications of mutation testing, e.g., equivalent mutant detection, automated test generation, and adaptation of mutation testing for machine learning-based systems.

Testing and Debugging AI Systems

AI Systems are computer programs that mimic the decision-making and problem-solving capabilities of the human mind. They generally consist of a machine learning component which operates based on rules deduced from training data and in turn increases the complexity of testing and debugging. In our research, we investigate how we can use conventional and unconventional testing and debugging methods to ensure the reliability, accuracy and trustworthiness of AI Systems

Collaborations

Diverse work of our group takes place in the context of collaborations. Under the umbrella of the Simulation Science Center Clausthal / Göttingen, we advance our software process simulation approaches and contribute to the deployment of simulations in the cloud with a Simulation Platform as a Service (SPaaS) approach. Furthermore, we are active in several Special Task Forces (STF) and in an Industry Specification Group (ISG) at ETSI. Here we contribute to the maintenance of TTCN-3, to the development of TDL, and to the specification of a framework for Augmented Reality applications. In the DFG funded GAIUS project we collaborate with the University of Greifswald on the development of the bioinformatics tool AUGUSTUS.