Git Modeling

Git Modeling

Rough explanation...

For all scientific modelling receiving public funding, and for all models whose results are used to sway public policy, all data needed for a model run, the code used in the model run, and the output from a model run should be put in a publicly readable git repository. This must be done for *every* run of the model.

We should be able to go back and rerun any model from any point in time with the data from that point in time and confirm that we get the same results that are published for that run in the past.

We should also be able to run data from point in the past A with the model code from point in the past B and compare the results to the results from data from point in the past A with the model code from point in the past C's results...

And other variations.