Heterogeneous experimental data in modern physics research facilities, such as KARA and FLUTE at KIT, often rely on experiment-specific metadata, making it challenging to define a unified schema for Research Data Management (RDM). We propose an object-oriented approach to address this challenge.In our framework, each experiment independently generates a dedicated class encapsulating its unique metadata. These classes are then collected and organized into a hierarchical tree structure using the concept of generalization, a core principle in object-oriented design. Shared properties across experiments are promoted to parent classes, while experiment-specific details remain in leaf classes.This method allows the automated generation of flexible and coherent metadata schemas for a wide variety of experiments, enabling interoperable, extensible, and maintainable RDM solutions. We demonstrate this approach using selected experiments at KIT, showing how class-based generalization produces a scalable and unified metadata representation, bridging the gap between heterogeneous experimental data and FAIR data principles.
Masoumi et al. (Tue,) studied this question.