基于网络的可信软件大规模协同开发与演化

随着网络时代的来临,软件的开发模式、运行环境和提供方式发生了巨大变化.互联网作为一种开放的协同工作环境,其中孕育的大规模协同创作机理对软件开发和应用产生了深刻影响.如何将其与工业化可信软件生产相结合以提高软件生产效率和质量,是网络时代软件技术面临的新课题.本文提出汇聚群体智慧的可信软件开发新方法——群体化方法,该方法的核心是"群体协同、资源分享、运行监控、可信分析",支持创新软件作品向可信软件产品转化,支持软件的可信演化.提出一种基于证据的可信软件概念模型,将软件在开发阶段、分享阶段和应用阶段的基础数据作为软件的可信证据,并据此建立了软件演化过程模型;提出一种支持可信软件协同开发与演化的服务模型,支持软件创作与生产深度协同的可信软件社会化生产、开放有序的可信软件资源分享,以及基于海量数据分析的软件可信评估.最后以"十一五"国家高技术研究发展计划重点项目"高可信软件生产工具及集成环境"为背景,阐述了面向群体化方法的关键技术、开发环境和应用实践.

Research on network-based large-scale collaborative development and evolution of trustworthy software

With the coming of the Internet era,great changes have taken place in the development models,runtime environments and providing styles of software. Being an open and collaborative working environment,Internet gives birth to mechanisms for successful crowd-based and collaborative creation,which have had pro- found impacts on development and application of software. How to combine these mechanisms with industrial approaches for trustworthy software production,and to improve the development efciency and quality of soft- ware,are new challenging issues of software technologies in Internet era. This paper proposes a new approach for trustworthy software development based on the crowd wisdom,namely Crowd-based Method,which centers on crowd collaboration,resource sharing,runtime monitoring and trustworthiness analysis,to support the transfor- mation from creative software works to mature software products,and supports the evolution of software. Firstly,an evidence-based concept model of trustworthy software is proposed,which classifes the data generated at the stages of software development,sharing and application as three kinds of trustworthiness evidence of software,and constructs a new process model for software evolution. Secondly,a service model for collaborative development and evolution of trustworthy software is proposed,which supports：（1） the socialized software development deeply combining the software creation and production;（2） the open and ranked trustworthy software resource sharing; and（3） the evaluation of software trustworthiness based on mass data analysis. Finally,with the background of the highly trustworthy software production tools and integrated environment（Trustie）,this paper expounds the key technologies,development environments and application practices for the proposed Crowd-based Method.