GazePrinter: Visualizing Expert Gaze to Guide Novices in a New Codebase

Why Read This from an HCI Perspective

From an HCI perspective, this article is worth reading because it shows how ‘gaze-based guidance’ can evolve from a mere tracking technique into a learning-support interface. In particular, it examines how visualizing experts’ exploration paths during onboarding to a new codebase affects novices’ reading strategies, cognitive load, and time efficiency. This makes it especially meaningful for researchers who are thinking about both developer-tool UX and learning-support design.

CIT's Commentary

From a CIT perspective, GazePrinter is an attempt to convey ‘expert knowledge’ not through explanations or comments, but at a more fine-grained level: attentional allocation. What’s interesting is that it can function both as an assistive mechanism for understanding code and as a stepping stone that helps novices internalize experts’ exploration rhythms. However, in real-world settings, expert gaze is not always the correct path, and depending on project context and task goals, it may even risk reinforcing or hardening bias. Therefore, this approach is likely to have stronger HCI value when designed as a ‘scaffolding for exploration’ rather than as ‘presenting the answer.’ CIT interprets key design variables for follow-up work as the visualization intensity, adaptive exposure, and how well users’ self-directed exploration is preserved.

Questions to Consider While Reading

• Q.When expert gaze visualization improves novices’ reading strategies, are there cases where it could instead reduce exploration diversity or reinforce incorrect paths?
• Q.How does GazePrinter’s effectiveness vary with task difficulty or the structural complexity of the codebase, and what would be a way to personalize it?
• Q.When deploying gaze-data-based guidance in a real IDE, how can we define the boundary between visualizations that reduce cognitive load and those that steal attention?