The Diagnostic Pathfinder has been used for nearly ten years at multiple colleges of veterinary medicine to teach diagnostic problem solving. A number of prior studies show this tool to be effective. Research in medical diagnostic problem solving provides hints, but no unambiguous answers regarding how such a tool should be designed. This in-depth review of the interface discusses each interaction in terms of how that interaction relates to the tool’s success. Nine faculty members who have taught using the Pathfinder during the last decade responded to interview questions regarding the tool. Their responses supported what had already been learned – that there is benefit when learner and instructor use the same process for solving a diagnostic problem, and then compare results, and when students learn in the context of realistic problems. Additionally, instructor responses suggest that the Pathfinder has been effective because it has 1.) enabled precise communication among experts and learners in a field where there is no generally agreed upon format for precisely communicating understandings of interrelationships between mechanisms of disease and clinical laboratory data, and 2.) provided a framework for manipulating data that respects the limitations of human memory and invites a thorough, explicit, and “artistic” rendering of the rationale.