Intermediate Model Specifications
LEMMA defines intermediate specifications for domain models, service models, and operation models in order to facilitate model processing. However, there exist no intermediate specifications for technology models and mapping models. First, technology information is a crucial concern in model processing. Therefore, LEMMA will directly incorporate possible technology information assigned to the elements of domain, service, and operation models into the corresponding intermediate models during the process of intermediate model derivation. Hence, model processors have immediate access to all technology information. Second, since mapping models import service models for their technology-specific refinement, the transformation of a mapping model into an intermediate representation boils down to the iterative derivation of the intermediate representations of all imported service models.
Benefits of Intermediate Models
While it would also be possible to leverage Xtext's standalone mechanism to parse LEMMA models and then iterate the resulting object graph for model processing purposes, the usage of intermediate models based on well-defined intermediate specifications is beneficial in model processing for several reasons:
- Intermediate representations of modeling language concepts can constitute a well-defined API for model processors.
- Intermediate specifications decouple modeling languages from model processors. As a result, modeling languages may evolve independently from model processors as long as the corresponding intermediate representations remain stable.
- Intermediate specifications decrease the semantic gap between modeling languages and certain kinds of model processing. That is, because they allow implicit incorporation of language-specific characteristics into intermediate models. Such characteristics may comprise, e.g., resolved inheritance hierarchies or reified default values. Consequently, technology-specific model processors like code generators can directly use these information. First, this form of preprocessing relieves model processors from acquiring language-specific information. Second, it fosters consistent behavior of model processors by centralizing the responsibility for the provisioning of language-specific information.
- In the context of LEMMA, intermediate models make model processing independent of a certain modeling technology. That is, because LEMMA serializes intermediate models in the vendor-independent XML Metadata Interchange (XMI) format.
What's Next?
In the following, we describe how to obtain the intermediate representation of a LEMMA model as an XMI file and provide the intermediate specifications for domain models, service models, and operation models. More precisely, these specifications constitute dedicated intermediate metamodels that govern the intermediate representations of LEMMA models both physically (i.e., within XMI files) and in-memory (i.e., as object graphs that originated from XMI parsing).