Scalable quality assessment for open-source solutions: An AI approach

In this presentation, we will show how generative AI (genAI) can be used to continuously measure various quality aspects in different development phases and derive targeted quality assurance measures. Using a case study from the smart city sector, we demonstrate the application of genAI for the evaluation of software repositories and the insights gained from it. The presentation will go into detail about the example to create a concrete understanding of the benefits of AI for software quality assessment. In addition, it provides a well-founded overview of how AI can contribute to improving software quality.

Value for the audience:
Using the software marketplace “Deutschland Digital” as a concrete example, the audience will gain a practical insight into how the measurement of the quality of open-source software through analyzing repositories and runtime artifacts (e.g. the actual implementation of user experience in user interfaces) with the help of generative artificial intelligence works. In concrete terms, the audience will take home successful approaches as well as lessons learned (e.g. approaches that have not worked) to be able to draw conclusions for their own software products and make the capabilities of generative artificial intelligence usable for their own organization.

Problems addressed:
The quality of software goes beyond the quality of static artifacts, such as code, and its measurement requires various quality metrics. How can these be implemented with genAI?

Talk language: English
Level: Advanced
Target group: Requirements Engineers, UX-Designers, Developers, Software Architects, Project Managers, Quality Managers

Company:
Fraunhofer-Institut für Experimentelles Software Engineering

Patrick Mennig

Frank Elberzhager