I am not a big fan of tracking mat files on git. I propose the following changes:
-
GUI Settings: Should not be tracked. These are personal and worst of all platform dependent. I've seen enough "C:" folders being created on linux when people run their code on the server without properly adjusting the output path.
-
Calibration files: In principle should also not be tracked. They are data dependent, and thus should be stored together with the data. Right now there are too many calibration files and some are obsolete. I admit it is somewhat useful to have them all at hand to track on someone else's data, but an effort should be made to store them somewhere else.
-
Model files: Only the model used on the paper should be tracked.
-
Other auxiliary files: If required by the system to work, then yes. If not, no.
I am not a big fan of tracking mat files on git. I propose the following changes:
GUI Settings: Should not be tracked. These are personal and worst of all platform dependent. I've seen enough "C:" folders being created on linux when people run their code on the server without properly adjusting the output path.
Calibration files: In principle should also not be tracked. They are data dependent, and thus should be stored together with the data. Right now there are too many calibration files and some are obsolete. I admit it is somewhat useful to have them all at hand to track on someone else's data, but an effort should be made to store them somewhere else.
Model files: Only the model used on the paper should be tracked.
Other auxiliary files: If required by the system to work, then yes. If not, no.