More Common Than You Think: An In-Depth Study of Casual Contributors

From AcaWiki
Jump to: navigation, search

Citation: Gustavo Pinto, Igor Steinmacher, Marco Aurélio Gerosa (2016) More Common Than You Think: An In-Depth Study of Casual Contributors.
Internet Archive Scholar (search for fulltext): More Common Than You Think: An In-Depth Study of Casual Contributors
Download: http://gustavopinto.org/lost+found/saner2016.pdf
Tagged: github (RSS), survey (RSS), open source (RSS)

Summary

"paper presents an empirical study aimed at illuminating the casual contributors [...] unique in its focus on understanding (1) how common they are, (2) what are the characteristics of their contributions, and (3) how they are perceived."

Study 1: "quantitative and qualitative analysis of data from GitHub"

Study 2: "two surveys conducted with 197 casual contributors and 65 project maintainers"

"RQ1. How common are casual contributors in OSS projects?"

"we found that a non-negligible number of contributors (48.98%) performed a single contribution. Based on this finding, we decided that the casual contributor is a contributor that performed at most one commit to a software project [...] these casual contributors are responsible for only 1.73% of the total number of contributions in our corpus of OSS projects."

"22.93% of the casual contributions changed a single line of code"

Found projects in dynamically typed languages have more casual contributions than statically typed ones.

"RQ2. What are the characteristics of a casual contribution?"

Identified 8 categories of casual contributions. Summary table:

  • 116 / 30% / Bug Fix
  • 110 / 29% / Documentation
  • 72 / 19% / Add New Feature
  • 34 / 9% / Refactoring
  • 25 / 7% / Update/Version Dependencies
  • 14 / 3% / Improve Error/Help Messages
  • 8 / 2% / Improve Resource Usage
  • 5 / 1% / Add Test Cases

"RQ3. How do casual contributors and project maintainers perceive casual contributions?"


"Most of the casual contributors and maintainers suggested that they are motivated by their personal needs (“scratching own itches”), that in most cases are related to fixing bugs that block or impact the development of other projects that depends on the analyzed projects. Yet, there are evidence that casual contributors do not become more active mainly because of time constraints. Finally, we found that although these contributions bring many benefits, they also cause some problems, mainly related to time required from core members reviewing the quality of code."

Theoretical and Practical Relevance

Ignores casual contributors who do not commit at all, eg those whole file issues. Counting anyone with more than one commit as non-casual seems a very low threshold.

Discussion at https://news.ycombinator.com/item?id=11977747