By Fernando Silva Parreiras, Jeff Z. Pan, and Uwe Assmann
Introduction
In the late 2000s, Model-Driven Engineering (MDE) was rapidly expanding, placing a heavy emphasis on model transformations, traceability, and the creation of Domain-Specific Languages (DSLs). However, as models became more complex, the software engineering community faced a "semantic gap": how to ensure model consistency, reason about them automatically, and formally represent the knowledge they contain. This document serves as the preface to the TWOMDE 2009 workshop, held in Denver in October 2009, which explored the intersection of MDE and Ontology development to address these challenges.
The Synergy of MDE and Ontologies
The authors argue that while Software Engineering and Ontology development (often associated with the Semantic Web) evolved separately, they are natural allies. The workshop aimed to bridge the gap between the modeling community and the knowledge representation community to improve how software systems are designed and verified. By leveraging semantic technologies—such as automated reasoning and formal knowledge representation—engineers can solve critical bottlenecks in software modeling and transformation.
Technical Bridges and Standardization
As a collaborative framework, the workshop identified several "tools of the trade" essential for this integration:
- Standardization Initiatives: Key contributions from the W3C’s Software Engineering Task Force (SETF) and the Object Management Group’s (OMG) Ontology Platform Special Interest Group (PSIG).
- Metamodeling: The use of the Ontology Definition Metamodel (ODM) and the OWL2 Metamodel as technical bridges. These allow ontology languages to be used within standard software modeling environments.
The workshop focused on three primary pillars:
- Applying scientific results from ontology research to MDE.
- Using ontologies to support and validate model transformations.
- Leveraging ontologies to improve the design and semantics of DSLs.
Key Insights: From Diagrams to Computable Models
A major insight from the workshop is that ontologies are not restricted to the "Semantic Web." Their core strengths—automated reasoning, dynamic classification, and consistency checking—are directly applicable to traditional software engineering tasks, such as ensuring that a model transformation preserves system logic.
The primary goal of integrating these fields is formalizing semantics in software. By using knowledge representation technologies, engineers can move from "informal" diagrams to "computable" models. This leads to the concept of Ontology-Driven Architecture (ODA), which guides software development through formal domain descriptions rather than purely code-centric approaches.
Results and Future Directions
The workshop successfully gathered researchers to identify success cases where integrated approaches combine MDE and ontologies. With concrete specifications like the OMG ODM now available, the field has reached a level of maturity suitable for broader adoption.
However, challenges remain in moving from theoretical success to widespread industry adoption. Future work must focus on:
- Practicality: Demonstrating value in mainstream industrial contexts.
- Tooling: Integrating reasoning engines directly into standard modeling IDEs.
- Education: Encouraging the modeling community to explore advanced ontology features, such as automated classification, which may still be unfamiliar to many software engineers.
Summary updated