Research On Ranking Opensource Software Based On Popularity

Software reuse is an important strategy for improving software development efficiency and increasing software systems quality.As one of the most important sources for software reuse,open source software(OSS)plays an essential role in promoting softwar reuse.Recently,with the development of open source movement,more and more developers are attracted by OSS communities.They take part in the process of developing OSS,sharing and learning knowledge of OSS.All those software resources and knowledge resources precipitat in the open source communities,accumulating a large amount of reusable resources.However,these massive open source resources are of large amount,various types,highly distributed,which results in great challenge for method of traditional method of lacating software resources for reuse.The main concerned problem of software locating is transforming from searching to ranking.In this paper,we make full use of the feedback information of the open source software users,and put forward a software ranking model based on popularity of software in communities.On the one hand,the vast amount of information about open OSS is often distributed in different communities.Users from open source communities ususlly develop software in software production communities,while discuss and learn knowledge in software consumption communities.All those make it a challenge to comprehensive understand software information.On the other hand,the discussion in open source communities is not static,and the relationship between the feedback and software is complex,which brings a challenge to the calculation software popularity.To face these challenges,this paper studies the following two questions:(1)build the connection among open source communities;(2)build the model of ranking OSS based on popularity.For the problem of building connection among open source communities,we propose a community connecting algorithm based on text matching to realize fast connecting information between different software production communitied and software consumption communities.Different from the method of subject identification,we make full use of resources' online attributes in open souece communities,such as description,tag,etc.to mining supplementary information.In the process of communities connecting,we carries on the connection and reducing duplication over 60000 software from 4 communities,and achieves a good accuracy and a fast effective connection resources form different open source communities.For the proble of building the model of ranking OSS based on popularity,we propose a ranking model based on discuss popularity of software in communities.Different from tranditional evaluating software method using software production and process data of development,we use users' discussion data to ranking software.In the process of building ranking model,we first get the popularity model of discuss through analyzing the changes of users' feedback.Then we use the relationship between software and user-feedback to build the communities popularity of software,getting the ranking model.Finally,we evaluate our ranking result with other communities' ranking result,and achieve a well ranking result.