The service provider must define and document how it keeps datasets accurate when they are built without manual paper entry, covering sets made by scanning, electronic or online collection, CAI databases and imports or merges of one source into another, synthetic data included. It must validate the automatic entry process at each project or stage and hold documentation testing both the design and the implementation, naming the tests used and the staff involved. Where scanning is used it must estimate rejection and substitution rates and record how rejected entries and substitutions are handled.
The graph holds this control, the 0 it maps to, and the evidence behind each claim, over MCP and REST.