Towards General Requirements for Norm Representation Languages Using Competency Questions: The Case of Comparing Flint and ODRL

The field of normative representation needs a way of evaluating whether representation languages are fit-for-purpose, and needs a way of understanding how representation languages can be compared or related to one another. We propose to use the methodology of competency questions, where model requirements are phrased as questions that should be answerable through the model’s data structure, as a basis for meta-level discussions about norm representation language design. To work towards this goal, we present an initial set of 52 competency questions intended for requirements testing of any norm representation language. We exemplify the usage of the competency questions by evaluating and comparing ODRL and Flint. We note that ODRL restricts itself to state-level information, while Flint additionally expresses state transitions; we identify competency questions that pinpoint these differences. We suggest that ODRL is primarily suitable for the purpose of policy enforcement, while Flint is better suited to the purpose of normative coordination between stakeholders. Thus, we demonstrate that subsets of the competency question set relate to various types of use cases, and we argue for a use case typology based on competency questions. We conclude that a widespread agreement on requirements, pursued through formulating competency questions, allows the study of normative representation languages to focus on its ultimate purpose: To understand and serve the variety of needs of all people who engage with norms.