Generative software module development for domain-driven design with annotation-based domain specific language

Abstract

Abstract Context Object-oriented domain-driven design (DDD) aims to iteratively develop software around a realistic model of the application domain, which both thoroughly captures the domain requirements and is technically feasible for implementation. The main focus of recent work in DDD has been on using a form of annotation-based domain specific language (aDSL), internal to an object-oriented programming language, to build the domain model. However, these work do not consider software modules as first-class objects and thus lack a method for their development. Objective In this paper, we tackle software module development with the DDD method by adopting a generative approach that uses aDSL. To achieve this, we first extend a previous work on module-based software architecture with three enhancements that make it amenable to generative development. We then treat module configurations as first-class objects and define an aDSL, named MCCL , to express module configuration classes. To improve productivity, we define function MCCGen to automatically generate each configuration class from the module’s domain class. Method We define our method as a refinement of an aDSL-based software development method from a previous work. We apply meta-modelling with UML/OCL to define MCCL and implement MCCL in a Java software framework. We evaluate the applicability of our method using a case study and formally define an evaluation framework for module generativity. We also analyse the correctness and performance of function MCCGen . Results MCCL is an aDSL for module configurations. Our evaluation shows MCCL is applicable to complex problem domains. Further, the MCCs and software modules can be generated with a high and quantifiable degree of automation. Conclusion Our method bridges an important gap in DDD with a software module development method that uses a novel aDSL with a module-based software architecture and a generative technique for module configuration.