Distribution of commits/changes per author over time about fornalder HOT 4 OPEN

Oh, that 's a good idea!

from fornalder.

We could also create cohorts based on the authors' all-time commit count. So instead of colors based on which year the members started interacting with the project, it'd be a color for 1-9 commits, another for 10-99, yet another for 100-999, etc. Then you could see when the big committers are more active compared to the smaller ones. These cohorts could be used in the active-contributors and commit-count charts, which would tell you e.g. "how many level 100-999, ... committers were active, per year" and "how many commits were made by level 10-99 committers, per year".

Part of why I'm suggesting this is also because it's easy to do -- add a new SQL query and reuse all the plotting and histogram code.

from fornalder.

We could also create cohorts based on the authors' all-time commit count.

I'm a bit unsure about using all-time commit count to identify "core contributors". One obvious issue would be cases where you have low level commits over a long period.

from fornalder.

I think someone doing polish/bugfixing or maintaining infrastructure consistently over a long time is likely important to the project. Although not as critical as someone speccing and writing the next version of GTK. I also received comments to the effect that measuring contributions by commits is unfair, from someone who spent a lot of time on analysis and API design that resulted in relatively few commits.

We should probably define what a "core developer" is.

from fornalder.