Avoid software conflicts during development

At the Brandenburg University of Technology Cottbus-Senftenberg (BTU), researcher Prof. Leen Lambers is developing a proactive approach for dealing with conflicts based on graph transformations in the new DFG project "Conflict Prevention and Proactive Conflict Resolution Strategies for Graph Transformation", or CORE-GT for short. The head of the Practical Computer Science / Software Systems Engineering chair is guided by the Capability Maturity Model Integration (CMMI), a process and maturity model that helps organisations to optimise processes and establish efficient ways of working in order to reduce risks in software, product and service development.

"The aim of mature processes is to prevent problems as far as possible instead of reacting to them after the fact," summarises Prof. Lambers. "So far, however, there has been little research into how conflicts in software development can be proactively prevented or resolved."

Graph transformation as a methodological basis

In Computer Science, graphs are a proven means of representation when complex relationships need to be illustrated or multi-layered structured data needs to be analysed. They can be used to visualise relationships between individual elements clearly and precisely, and the concepts of graph transformation provide a powerful basis for developing an effective approach.

According to Prof. Lambers, there are three reasons for this:

• Graphs are well suited to representing complex structures, such as code artefacts, System Architecture or computational states.
• Changes to these structures can be clearly described using graph transformation rules.
• Graph transformation is based on a rich and well-established theory.

From prevention to solution

Potential conflicts should be avoided during development with the help of prepared repair proposals. If they cannot be prevented, pre-defined solution templates are used, which are applied semi-automatically if necessary and orientated towards a solution that is viable for both sides using heuristic search procedures.

The CORE-GT project is led by Prof. Gabriele Taentzer (Philipps University Marburg) and Prof. Leen Lambers. Two positions are earmarked for realising the project ideas within three years and writing doctoral theses.