Where the test sequence reveals a failure, or a likely future failure, that makes repairs or modifications necessary, the laboratory and the supplier agree either on a new sample on which every test influencing the result is repeated or, preferably, on making the repairs or modifications and repeating only the relevant tests.
The graph holds this control, the 0 it maps to, and the evidence behind each claim, over MCP and REST.