摘要
Recently software crowdsourcing has become a viable development paradigm for Software-as-a-Service （SaaS） ecosystems. TopCoder, one of the largest competi- tive programming communities, enables enterprises to tap into its global talent pool and crowdsource a variety of SaaS development tasks including requirement analysis, architec- ture design, code, and testing. Many researchers have pro- posed auction-based modelling methods to characterize gen- eral software crowdsourcing. But there are few papers on the comprehensive analysis of SaaS crowdsourcing process and developer community. This paper introduces a holistic analysis framework to model the SaaS-oriented software crowdsourcing from two dimensions： individual behavior in crowdsourcing contests and collective competition in the community. The framework includes a game-theoretical model to describe the compet- itive nature of software crowdsourcing process. Moreover, the framework defines a competition network to characterize the topological properties of a crowdsourcing community for SaaS development. The analysis of this model indicates that the success of a competitive software crowdsourcing project essentially depends upon the networks of key participants with sufficient skills and dedication for the project. This is validated by a large historical data collected from the Top- Coder website over a ten-year period.
Recently software crowdsourcing has become a viable development paradigm for Software-as-a-Service （SaaS） ecosystems. TopCoder, one of the largest competi- tive programming communities, enables enterprises to tap into its global talent pool and crowdsource a variety of SaaS development tasks including requirement analysis, architec- ture design, code, and testing. Many researchers have pro- posed auction-based modelling methods to characterize gen- eral software crowdsourcing. But there are few papers on the comprehensive analysis of SaaS crowdsourcing process and developer community. This paper introduces a holistic analysis framework to model the SaaS-oriented software crowdsourcing from two dimensions： individual behavior in crowdsourcing contests and collective competition in the community. The framework includes a game-theoretical model to describe the compet- itive nature of software crowdsourcing process. Moreover, the framework defines a competition network to characterize the topological properties of a crowdsourcing community for SaaS development. The analysis of this model indicates that the success of a competitive software crowdsourcing project essentially depends upon the networks of key participants with sufficient skills and dedication for the project. This is validated by a large historical data collected from the Top- Coder website over a ten-year period.
