Engineering a cognition-based specification method - Software and Systems Modeling

Context Software development is inherently a human cognitive task that involves the capture and integration of diverse knowledge and decisions from multiple stakeholders. Existing specification methods and languages mostly rely on computer-based or mathematical primitives, leading to a disconnect between how people naturally think and communicate, and how systems are specified. Therefore, we are investigating a method, called MuDForM (Multi-Domain Formalization Method), to formalize and integrate the knowledge of multiple domains into domain models and into specifications in terms of those domain models. We created a first coherent definition of the method, which emerged from several case study evaluations published in previous works. Goal Establish a method definition that is explicitly based on concepts from human cognition. Method We studied literature in (language) philosophy, linguistics, and cognitive science, to identify concepts that can serve as the cognitive underpinning of the method’s metamodel. We made a model of those concepts and connected them to MuDForM’s metamodel via an explicit method definition structure and analyzed the result. Result The paper defines the conceptual and structural groundwork for illustrating how a specification method can be constructed based on insights into human cognition and communication. It provides a coherent model of cognitive specification aspects, which grounds the modeling concepts in MuDForM ’s metamodel and makes it cognition-based. We also identified additional cognitive aspects that call for future work and the possible extension of the metamodel. The paper clarifies MuDForM ’s objective to support the transformation of natural language into unambiguous, cognition-aligned models.