Guiding Software Engineers by Eye-Tracking (EyeGuide)

• Applicant: Professor Dr. Kurt Schneider
• Subject Area: Software Engineering and Programming Languages
• Term: since 2025
• Project identifier: Deutsche Forschungsgemeinschaft (DFG) - Project number 560847561

Many software development tasks require collaborative work. Requirements engineers, designers, and other software professionals need to gain an overview of a domain and a project quickly, consider previous work and relationships between different documents, and on-board quickly to a new project or keep up-to-date with an on-going project. It will be crucial to understand large amounts of information during the entire project. It would be good to be guided through the heap of documents by someone who has explored it before, but such a guide causes even more effort and will hardly be affordable. Therefore, after one person has invested this time and effort, the next person needs to start over again, reading and making sense of documents. This is neither efficient nor effective. Automated support for guiding software developers is missing. In EyeGuide, we propose to use eye-tracking in an operational way to provide guidance: (1) To familiarize oneself with documents and artifacts effectively, (2) to support recognizing dependencies between them for understanding the bigger picture faster and better, (3) to catch up on the discussion and status of the project team, and, finally, (4) to develop a shared understanding quickly and facilitate collaboration – without a lot of extra time or work needed. In this project, we propose to share attention by eye-tracking between software professionals. Eyetracking delivers traces of attention (eye-traces). Rich common ground has been demonstrated to improve speed and quality of collaboration. Learning how much attention was devoted to different artifacts and their parts or sections can help to conclude what is difficult to understand, what are key achievements, and what should be read first or at high priority. We use eye-tracking as an interactive medium, not just for scientific analysis and insights. This proposal suggests using eye-traces for guiding developers and project. We focus on eye-tracking since it is a unique technique to enable recognition and capturing of areas, durations, and patterns of reading and attention. The expected benefit is to conduct software tasks faster, more competently, and with better result. Eye movement is used like a pointer to highlight and profile a given set of documents. Collected eye-tracking data is analyzed, aggregated, and transformed to a useful format for recipients. The resulting representation can guide software engineers to gain shared understanding faster. They can focus on the most important parts faster and invest their time and resources more effectively.

• DFG Programme: Research Grants