The operator puts in place controls, matched to the level of risk, that remove public risk so far as is reasonably practicable or, where removal is not reasonably practicable, reduce it so far as is reasonably practicable, choosing measures that work in its real operating conditions rather than from generic lists; controls may touch equipment, vehicles, loads, scheduling, procedures, resourcing, roles or how work is organised, and no code, hierarchy, tool or paperwork is mandated. Seven things are looked for. Controls are chosen to eliminate risk where reasonably practicable and otherwise to minimise it. Each control matches the nature and level of the risk and can deliver the intended safety outcome. Controls suit the operator's environment, activities, interfaces and conditions rather than assumed solutions. Where one control is not enough, several are layered to reduce risk effectively. Controls are actually applied in daily work rather than living only in documents. Controls get the support they need to function: communication, information, coordination with other parties, resources. And where a control fails to manage a risk adequately, alternatives or additions are considered and put in place.
The graph holds this control, the 0 it maps to, and the evidence behind each claim, over MCP and REST.