From Ontology to Domain-Specific Languages: Turning Domain Expertise into Actionable Models

This talk describes how to create domain-specific languages that are directly based on knowledge expressed in ontologies, glossaries, and terminologies.

Such domain-specific modeling languages offer several benefits:

1. They allow domain experts and subject matter experts to express their knowledge directly using familiar and well-known terms. There is no need to map existing domain terms to other languages or learn new modeling languages. Using already known terminology makes the transition to model-based engineering easier.

2. They guide modelers to specify knowledge in a consistent and complete manner, producing precise specification models.

3. Models that are precise and high-fidelity can be used for checking, tests, configuration, and generating software code automatically.

Together, these benefits lead to improved quality and increased development productivity.

During the talk, we present examples of domain-specific languages from both systems and software development.

The main part of the talk focuses on turning domain knowledge and ontology definitions into a working domain-specific modeling language. We describe what additional elements are needed beyond glossaries and ontology definitions, such as:

1. Constraints and rules for modeling;

2. Concrete syntax, such as diagrams, maps, matrices, tables, etc., for representing models to humans;

3. Generators to produce code, tests, reports, etc. from the models.

We demonstrate language creation with practical examples and present cases where domain-specific languages are used to create and test features, as well as to generate running software code and system deployment instructions from models.

We conclude by discussing the role of robust tooling that allows ontology definitions and corresponding languages to evolve without breaking existing models or editors.