The processing code is used in the PROCSPEC.CODE record field. It specifies how subsets of the data have been processed. This code is different from the task identifier, which identifies a task that was used to produce the dataset. Instead, the processing code provides additional information about the data itself; think of it as a sort of ``Post-It'' note attached to some subset of the data. For example, the lead time of forecasts, the interpolation parameters for subsets of the data, and parameters used in a model run can all be specified here. These codes are integer values at least 32 bits long.