Towards dynamic classification in domain modeling with Jjodel

Ter Beek Maurice Henri, Bucchiarone Antonio, Pierantonio Alfonso, Selic Bran

Dynamic Classification; Domain modeling; Evolving systems; Jjodel 

This work addresses the limitations of traditional object-oriented classification in representing evolving systems. Conventional classification enforces rigid hierarchies that hinder dynamic reclassification, concurrent viewpoints, and transient or overlapping states, thus limiting their usefulness for systems that exhibit change, context sensitivity, or adaptation. We propose a modeling notation that extends UML class diagrams with declarative behavioral dynamics directly embedded in structural information. Unlike approaches that separate structure and behavior, our notation unifies them in a framework suitable for both conceptual/domain modeling and evolutionary, context-aware dynamics. The notation is fully defined, including abstract syntax, diagrammatic syntax, semantics, and simulation, within the Jjodel platform, ensuring rigor and tool support. Its structural nature enables the declarative specification of dynamics without imperative constructs. Moreover, the notation can be connected to frameworks for formal verification and model checking, enabling analysis of dynamic properties. Applicability is illustrated through a context-aware scenario in which enriched structural models can be simulated, reasoned about, and eventually verified.

Source: LECTURE NOTES IN COMPUTER SCIENCE, vol. 16470, pp. 193-215

Publisher: Springer

Metrics