Avionics systems are becoming more and more complex due to the demand of additional functionality with very high safety and reliability needs, as well as the emergence of novel digital technologies. This leads to an almost unmanageable effort required to develop avionics systems, which are characterized by their numerous parameters and relationships. Therefore, conventional development methods are on the edge of economic feasibility, leading to a stronger reliance on new development methodologies such as model-based systems engineering (MBSE). MBSE enables the structured representation of complex systems using domain-specific modeling languages (DSMLs), which serve as blueprints for user models—digital representations of real-world systems. DSMLs map a respective domain with its specific properties and allow to keep the language of that domain.

One major challenge is the vast number of available language workbenches for DSMLs, each differing in design, features, terminology, and application areas. This makes it difficult for developers to understand overarching principles and design alternatives. To address this issue, we have developed language-independent commands for model interaction based on commonly used modeling concepts. A “concept” represents a specific modeling aspect with a defined meaning—for example, the concept M2CLASS refers to a typical class element that serves as a blueprint for instantiating objects at the M1 level. These concepts can be used independently of the meta-modeling language or workbench, offering greater flexibility for developers.

Moreover, these concepts can be utilized as a novel method for text-based model persistence. Each model element is uniquely defined through a sequence of creation commands that establish the model structure and its associated information, such as class names or attribute assignments, which are processed by our model interaction language.

Managing DSML changes over time remains a major challenge. Modifying DSMLs can lead to incompatibilities, potentially rendering “outdated” user models inoperable, loosing data or unintentional modification, especially in avionics development, where multiple stakeholders collaborate. To maintain a consistent and reliable development process, these incompatibilities must be addressed systematically. Our “concept” based persistency could be utilized for the handling of DSML changes by simply comparing the persistent text-based model storage. The strongly defined structure enables the derivation of a closed set of modification types. To manage those changes, they can be categorized into the following three types:

  1. Non-breaking changes, which do not affect existing user models.
  2. Breaking-resolvable changes, which impact user models but can be managed automatically.
  3. Breaking-nonresolvable changes, which require manual intervention to update the user model.

This talk will present the motivation behind our concept-based modeling approach and provide insights into its current development stage with a focus on the model changes.