We propose a formally defined method to transfer the core of class-like models lossless between domain-specific frameworks and evaluate that for two frameworks. By establishing common modeling concepts and a command interface, we wrap structurally different meta-modeling approaches to exhibit bisimilar behavior for meta-models (M2) and user models (M1) across frameworks. Modeling approaches share concepts such as instantiation and inheritance, but exhibit different syntax and semantics. Porting models from one domain-specific modeling framework to another requires dedicated conversions and, after porting, working with the model can be different. We identified 17 common concepts and features. Using these concepts, we derived a framework-independent model interaction interface based on CRUD commands. This interface allows the creation of M2 and M1 models without worrying about the individual syntax and semantics of a framework. M1 and M2 models can be queried and edited with identical commands and the results are identical for different underlying domain-specific frameworks. An analogy of labeled state transition systems (LSTS) and our interface prevents model inconsistencies and ensures safe continuous model modifications. Moreover, comprehensive tests can be derived that verify that a wrapped framework simulates our interface and, thus, two wrappers form a bisimulation as long as dedicated framework assumptions are valid. In a case study, we evaluate the effort and practitcability of wrappers for pyecore and adaeoq. The results demonstrate the feasibility and show that meta-models and user models with thousands of elements behave bisimilar while the implementation effort is acceptable.

The development of cyber-physical systems can significantly benefit from domain-specific modeling and requires adequate (meta)-modeling frameworks. If such systems are designed for the safety-critical area, the systems must undergo qualification processes defined and monitored by a certification authority. To use the resulting artifacts of modeling tools without further qualification activities, the modeling tool must be additionally qualified. Tool qualification has to be conducted by the tool user and can be assisted by the tool developer by providing qualification artifacts. However, state-of-the-art domain-specific modeling frameworks barely support the user in the qualification process, which results in an extensive manual effort. To reduce this effort and to avoid modeling constructs that can hardly be implemented in a qualifiable way, we propose the development of an open source (meta)-modeling framework that inherently considers qualification issues. Based on the functionality of existing frameworks, we have identified components that necessarily need to be rethought or changed. This leads to the consideration of the following six cornerstones for our framework: (1) an essential meta-language, (2) a minimal runtime, (3) deterministic transformations, (4) a qualification artifact generation, (5) a sophisticated visualization, and (6) a decoupled interaction of framework components. All these cornerstones consider the aspect of a safety-critical (meta)-modeling framework in their own manner. This combination leads to a holistic framework usable in the safety-critical system development domain.