GrimoireLab & Crossminer: Two different philosophies to measure open source community health

Decision makers often face challenges to understand open source projects, especially as data is scattered across different collaboration platforms. Two tools - GrimoireLab and CROSSMINER - were developed to overcome this challenge but were started with different goals in mind. On the one hand, GrimoireLab focused on understanding software development process within open source projects. On the other hand, CROSSMINER focused on helping software developers select the right open source projects. The meaning of project health is different in both cases. This talk is about this philosophical difference in approaching open source project health metrics, how it shaped the tools, what this means for metrics that were implemented, and what lessons can be learned about measuring open source project health.

Transcript

1. GrimoireLab & CROSSMINER Two different philosophies to measure open source community health {valcos, georglink}@bitergia.com @_valcos_, @GeorgLink CHAOSScon NA, 20th August 2019
2. /anecdote Car, Bus, Train, or Plane? ✈
3. /the question How can we know if an open source community is healthy?
4. /community health Processes Activity Code quality Community Licensing Documentation Popularity
5. /community tools
6. /diff. philosophies Is a project community healthy? Which is the healthiest projects to adopt?
7. /grimoirelab
8. /grimoirelab chaoss.biterg.io
9. /grimoirelab chaoss.biterg.io
10. /crossminer
11. /crossminer scava-dev.ow2.org:8601
12. /crossminer scava-dev.ow2.org:8601
13. /crossminer
14. /comparison • Community metrics • Fine-grained data • Reactive metric calculation • Multi project aggregation • Code quality metrics • Coarse-grained data • Pre-determined, daily metric calculation • Mono project aggregation • Quality models
15. /comparison • Fine-grained data • Reactive metric calculation • Community metrics • Multi project aggregation • Coarse-grained data • Fixed metric calculation (daily) • Code quality metrics • Mono project aggregation • Quality models A quality model simplifies comparing many projects across different indicators of community health.
16. /lessons learned Has multiple meanings Shapes the indicators and metrics Influences technical implementations Can be summarized by a quality model Community Health
17. /discussion Your reaction? What philosophies have you seen? CHAOSS metric standard across software? Metric implementation transfer? Re-use of data in different designs?
18. Valerio Cosentino, Georg Link {valcos, georglink}@bitergia.com @_valcos_, @GeorgLink Thank you! Enjoy CHAOSScon! @grimoirelab @bitergia @CHAOSSproj @crossminer