Unit of (numeric) data and accompanying metadata

Every measurement (or calculation) produces (raw) data that are useless without additional information, such as experimental parameters. This additional information is termed metadata. A dataset is the unit of (numerical) data and metadata. Another integral aspect is the history containing all relevant information regarding each single processing step performed on the data of the dataset.

Reproducibility is an essential aspect of good scientific practice. In the context of data processing and analysis, this means that each processing step performed on data (of a dataset) should be stored in an reproducible way and preferably in a consistent format.

To be of actual use, an entry of the history needs to contain all information necessary to reproduce the processing step in its original form. This includes as a minimum the name of the processing routine used, the complete list of necessary parameters for that routine, and a unique version information of the routine. Additional useful aspects contain information about the operating system used, the name of the operator, and the date the processing step has been performed.