Off to a Good Start: Dynamic Contribution Patterns and Technical Success in an OSS Newcomer’s Early Career

Abstract

Attracting and retaining newcomers are critical aspects for OSS projects, as such projects rely on newcomers' sustainable contributions. Considerable effort has been made to help newcomers by identifying and overcoming the barriers during the onboarding process. However, most newcomers eventually fail and drop out of their projects even after successful onboarding. Meanwhile, it has been long known that individuals' early career stages profoundly impact their long-term career success. However, newcomers' early careers are less investigated in SE research. In this paper, we sought to develop an empirical understanding of the relationships between newcomers' dynamic contribution patterns in their early careers and their technical success. To achieve this goal, we compiled a dataset of newcomers' contribution data from 54 large OSS projects under three different ecosystems and analyzed it with time series analysis and other statistical analysis techniques. Our analyses yield rich findings. The correlations between several contribution patterns and technical success were identified. In general, being consistent and persistent in newcomers' early careers is positively associated with their technical success. While these correlations generally hold in all three ecosystems, we observed some differences in detailed contribution patterns correlated with technical success across ecosystems. In addition, we performed a case study to investigate whether another type of contributions, i.e., documentation contribution, could potentially have positive correlations with newcomers' technical success. We discussed the implications and summarized practical recommendations to OSS newcomers. The insights gained from this work demonstrated the necessity of extending the focus of research and practice to newcomers' early careers and hence shed light on future research in this direction.

On the sustainability of deep learning projects: Maintainers' perspective

“…For OSS projects, we adopted the 20 OSS projects published in Wang et al's 23 study and further obtained the top 10 most starred projects that come from Apache and OpenStack ecosystems, respectively. The rationale why we further select these two ecosystems is that they are representative in traditional OSS domains and are widely investigated by many researches 24,25 . As a result, we obtained 40 traditional OSS projects.…”

“…The rationale why we further select these two ecosystems is that they are representative in traditional OSS domains and are widely investigated by many researches. 24,25 As a result, we obtained 40 traditional OSS projects. To safeguard the quality of the dataset, we conducted the filtering rules applied to DL projects again and further selected traditional OSS projects with a wide range of tech stacks and application domains, such as a web development framework (react), a team collaboration tool (zulip), and a database (tidb).…”

“…To alleviate this threat, we referred to extant literature and found that Wang et al's 23 study only collected 20 OSS projects to unveil elite developers' activities, and Joblin et al 35 obtained 32 OSS projects to understand how socio‐technical factors affect the success of projects. Moreover, Yue et al's 25 study selected 54 OSS projects to investigate the correlations between newcomers' dynamic contribution patterns with their technical success in their early careers. Our constructed dataset includes 38 OSS projects, which is quite similar in scale with prior studies.…”

On the Sustainability of Deep Learning Projects: Maintainers’ Perspective

“…For OSS projects, we adopted the 20 OSS projects published in Wang et al's 23 study, and further obtained the top 10 most starred projects that come from Apache and OpenStack ecosystems, respectively. The rationale why we further select these two ecosystems is that they are representative in traditional OSS domains and are widely investigated by many researches 24,25 . As a result, we obtained 40 traditional OSS projects.…”

“…Moreover, Yue et al's 25 study selected 54 OSS projects to investigate the correlations between newcomers' dynamic contribution patterns with their technical success in their early careers. Our constructed dataset includes 38 OSS projects, which is quite similar in scale with prior studies.…”