Using the Skill–Rule–Knowledge (SRK) Lens for Analyzing Human Cognitive Modes and their Transitions in Software Engineering

Context: Software engineering (SE) is inherently cognitive, yet much of existing research and practice treats human cognition as a uniform and stable factor. This simplification limits our ability to explain key phenomena such as performance variability, inconsistent effectiveness of processes and tools, and mixed outcomes of AI assistance. Objective: This paper encourages the adoption of the Skill–Rule–Knowledge (SRK) model as a lens for cognitive control in SE, to enable a more precise analysis of cognitive modes and their transitions during SE activities. Method: We draw on the SRK model from human factors literature and reinterpret SE activities across the software development lifecycle in terms of skill-based, rule-based, and knowledge-based modes. Through analytical synthesis, illustrative scenarios, and an initial empirical assessment — comprising practitioner surveys and interviews, project artifact analysis, and evaluation of SRK-informed reformulations of existing SE research questions — we examine how cognitive modes and their transitions shape SE work. Results: Our analysis and empirical assessment show that making cognitive modes explicit reveals insights that remain obscured under simplified cognitive assumptions. Practitioners identified analytical value in SRK distinctions that mode-blind accounts would miss, project artifacts confirmed that issues tend to cluster at cognitive mode transition points, and SRK-informed reformulations of SE research questions were rated as more precise and insightful than their originals. Conclusion: The SRK model can provide a useful analytical lens for studying SE work at a finer granularity. It is not intended as a prescriptive model for individual engineers, but it can offer value for researchers and senior practitioners. This paper could lay the groundwork for further SRK-informed empirical and analytical research in SE.