IEEE 7000 9.3(c): Formulate and record the value-based system requirements arising from each EVR

The project shall study the risk mitigations and value demonstrators contained in the EVRs to find candidate value dispositions; formulate for each EVR one or more value-based system requirements (functional or otherwise) that make it real within the system (for machine learning systems these can include data collection, formats, diversity, ranges, provenance, performance measures, explainability, fairness evidence and training-data rules; for organizational systems personas, capability analysis and process models); set acceptance criteria and measurement targets, qualitative or quantitative, for every requirement; and record each with a unique reference, its traceability to an EVR, its risks, assumptions and constraints, noting the risk of acceptance criteria that cannot be made measurable.