A provenance-based solution for software selection in scientific software sharing

Xing Huang, Tun Lu, Xianghua Ding, Tiejiang Liu, Ning Gu

Abstract: As a form of collaborative work, sharing plays an important role in scientific research. Scientists share their discoveries with others by the ways of publishing paper, data and software. It promotes the development of research. With the increasing number of scientific software and more and more open source communities springing up, software selection becomes an urgent issue in sharing. In this paper, from the users' perspectives, we first propose a collaborative provenance model, which is designed to record the collaborative information between users and software. Then, drawing on ethnographic method, we explore the elements influencing the choice of users, and quantify them. On that basis, an algorithm is introduced to select the software for each process in workflow. For the whole workflow, a composition selection algorithm is presented.