![]() |
||||||||||||||||||||
|
||||||||||||||||||||
| REQUIREMENTS | ARCHITECTURE | IMPLEMENTATION | TRACEABILITY | CASE STUDIES | PUBLICATIONS | ||||||||||||||||||||
|
Documents -> Implementation
Aspect-Oriented, Model-Driven Implementation Techniques for Software Product Lines
In the project we investigated advantages of existing aspect-oriented and model-driven techniques for variation management in product lines and proposed a set of ideas for their further improvements. The results of this analysis and the improvement proposals can be found in the deliverables D3.1 and D3.2. The improved implementation platform for SPLs, developed in the context of this project, consists of a MDD toolsuite openArchitectureWare and an aspect-oriented language ECaesarJ. In the context of the project, the openArchitectureWare toolsuite was extended with aspect-oriented techniques in order to enable aspects in all phases of a MDD process. Aspects at the level of models are enabled by the XWeave tool, whereas the aspect-oriented extensions to the transformation languages of oAW, Xpand and Xtend, enable aspect-oriented decomposition of model-to-model and model-to-text transformations. The aspect-oriented techniques in the MDD process enable modularization of variations at different levels of abstraction. Figure 1. Integration of AOP into different phases of a MDD process
The ECaesarJ language enables flexible modularization of feature-oriented variation at the level of implementation. The flexibility of decomposition is achieved by supporting polymorphism and composition semantics for a broad set of software entities: methods, classes, events, and state machines. All these software entities can be declared as abstract and be used to describe explicit interfaces between implementations of different features of a SPL. Consequently, the language improves the balance between flexibility of software decomposition and encapsulation of resulting components, which facilitates better modularization of SPL implementation.
|
||||||||||||||||||||
|
© 2007 AMPLE |