Some improvements we should probably make to code credit (fairly easy):
[ ] Initial commits / early days of direct commit pushing should be accounted for. Only PR diffs are used at the moment. We could create a prs/0.json for example where we look at commits prior to, say, Aug 25, 2011 and credit existing code that way. And whatever GitHub says, I recall that fwd/inv/fiff code at that point should be credited to at least Alex Gramfort and Matti Hämäläinen.
[ ] We could consider adding contributor counts to the modules, like mne (123), mne.cov (45) or whatever the counts actually are.
Some improvements to credit beyond code (will be more work):
[ ] Add something having to do with PR comments, and issue comments. Not sure if we want to use number of comments, word count, etc. All of these metrics are flawed in different ways, so not sure the best thing here.
[ ] Add something about forum posts. This will require some discourse querying.
Just getting some ideas down that have been rattling around in my head, happy to add more ideas.
Some improvements we should probably make to code credit (fairly easy):
diffs
are used at the moment. We could create aprs/0.json
for example where we look at commits prior to, say, Aug 25, 2011 and credit existing code that way. And whatever GitHub says, I recall that fwd/inv/fiff code at that point should be credited to at least Alex Gramfort and Matti Hämäläinen.mne (123)
,mne.cov (45)
or whatever the counts actually are.Some improvements to credit beyond code (will be more work):
discourse
querying.Just getting some ideas down that have been rattling around in my head, happy to add more ideas.