Towards Combining Ontologies and Model Weaving for the Evolution of Requirements Models

Software change resulting from new requirements, environmental modifications, and error detection creates numerous challenges for the maintenance of software products. While many software evolution strategies focus on code-to-modeling language analysis, few address software evolution at higher abstraction levels. Most lack the flexibility to incorporate multiple modeling languages. Not many consider the integration and reuse of domain knowledge with design knowledge. We address these challenges by combining ontologies and model weaving to assist in software evolution of abstract artifacts. Our goals are to: recover high-level artifacts such as requirements and design models defined using a variety of software modeling languages; simplify modification of those models; reuse software design and domain knowledge contained within models; and integrate those models with enhancements via a novel combination of ontological and model weaving concepts. Additional benefits to design recovery and software evolution include detecting high-level dependencies and identifying differences between evolved software and initial specifications.