Closed jesparent closed 1 year ago
I had a meeting with Himanshu regarding his last year’s project and it’s implementation.
He told that it was basically an attempt to demonstrate the phenomena of a contributor being promoted to a maintainer by solving issues and increasing the contributor's confidence points.
This is a simulation of the model : -
As you can see in the simulation, there are 3 types of issues - earth-shot
, moon-shot
and mars-jupiter shot
. They have been categorised according to their difficulty levels (increasing order).
A contributor receives 10, 20 and 50 confidence points for solving a earth-shot
, moon-shot
and mars-jupiter shot
issue respectively.
Once a contributor attains 90 confidence points, he is promoted to a maintainer.
This model makes use of the Netlogo's Q-learning plugin.
Himanshu has suggested me to get a bit comfortable with RL, then try to create these models in Mesa using RL. Once this stage is reached, further work can be done on the model.
nice - let's think about what the next steps are and make specific plans @rv602 @Orthogonal-Research-Lab
Detail / report back on meeting with Himanshu