基于特征精化的产生式编程方法研究

Feature Refinement-based Generative Programming Method

针对当前产生式编程方法不能有效地应用于复杂软件系统(如信息系统)的开发,不能在分析和设计的层次上复用软件等问题,本文提出了一种基于特征精化的产生式编程方法,基本思想是首先使用特征模型(Feature Model)描述领域内的概念和特征,然后使用特征精化的方法将特征精化为基本特征,用基本特征和之间的关系说明特征是如何实现的,最后将基本特征映射为组件,根据系统的特征模型组装组件,构建整个软件系统。基于特征精化的产生式编程方法是建立在特征和组件的基础上的,适用于信息系统类软件的开发。

This paper presents an efficient Generative Programming （GP） method based on feature refinement and components for development of complex software system such as information system. The basic idea is to present domain features and concepts using the form of features and feature models, which can describe differences and commons of software on level of analysis and design. Then features are refined to basic features, which present internal structure of features. Finally, basic features are mapped to component or components. According to the feature models of system, software is generated through compositing components automatically by generator. This method is suitable for the development of information system software.