Models within DSSAT are developed collaboratively with input from people with expertise ranging from crop physiology, to soil science to software engineering. Typically, one or more people identify features they would like to see added or issues they feel need to be fixed. They contact the core development team and a roadmap is produced. If needed, scientists are provided assistance in configuring software to edit the FORTRAN code and manage code changes via GitHub, usually by attending a Development Sprint. Depending on the nature of changes, additional data sets may be sought for model calibration and testing. When the modifications are stable, they are shared via GitHub for additional testing. A recurring concern is that modifications that improve performance for one situation (e.g., set of experiments from a region) do not adversely affect other simulations. In certain circumstances, allowance for alternative calculations can be provided via flags that a user can set in the Simulation Control section of the crop management file.