Within Open Source projects, it is often useful to get a level set" on community practises (what works and what doesn't). The purpose of this talk will be to compare and contrast the Current OpenStack community processes with those of the Linux Kernel (being one of the oldest and best established Open Source projects with one of the biggest contributor bases). The object isn't a winner takes all competition but an in depth comparison of the two project on a set of metrics to see which may be better, what may be learned (and perhaps how they could both improved). It is important to understand as well, that there are often no "right" answers. Sometimes the dynamics of a community become a matter of taste within that community and a process that works perfectly for one community may fail dismally for another.
Some of the main measures of health of an Open Source project are
- Ease of attraction of newcomers
- Utility of the project processes to drive by coders (people who are only interested in getting one bug fixed or feature accepted and then they disappear again)
- Core contributor variation with time (who are the core contributors, how steady are they and what's the amateur/professional mix)
However, there are also other interesting issues, like what are the barriers to contribution, like how easy is the Contributor Licence Agreement to actually agree to (here the Kernel wins because the Developer Certificate of origin is streamlined and carried within the patch itself) and how welcoming is the community (here OpenStack wins, because the Linux Kernel mailing list is notoriously hostile); how does the acceptance process contribute to the quality of the code (OpenStack has a formalised gerrit based review process and the Kernel just has mailing list based discussion of patches).
The presentation will attempt to synthesise the community and process comparisons and shed some light on what is currently working well, and what may benefit from modification (and how might it be modified).