6.10 Safety measures

What this heading covers

Safety here is about harm to people, property or the environment from the product in its context, not about the absence of defects. The five families cover the constraints the product must operate within, its ability to identify risks, its behaviour when it fails, the warnings it gives, and its integration with other systems in a way that stays safe. This is the characteristic most likely to be new to a team whose measure library follows the earlier eight-characteristic model, and the one where a product quality measure can overlap with a sector safety case. The draft keeps the measure at the product level; it does not replace a hazard analysis or a sector regime. The 2023 model also adds intervenability under safety, and the draft gives it no separate family.

Record position: characteristic 9 of 9, carrying 5 of the 40 measure families in clause 6.

What the heading asks of a team

The common confusion

Safety is new to the product quality model, and it is often confused with the absence of defects. It is about harm in context: constraints, risk identification, safe failure, warning and integration.

The measure families in this characteristic

• 6.10.2 Operational constraint measures
• 6.10.3 Risk identification measures
• 6.10.4 Fail safe measures
• 6.10.5 Hazard warning measures
• 6.10.6 Safe integration measures

What to put in place before publication

1. Name the harm the product could cause and the context in which it could occur, before choosing any measure.
2. Write the operational constraints the product must respect, and the measure that shows it does.
3. Define the safe state for the product and the conditions that should trigger it.
4. Record how hazards are identified and how warnings are delivered and acknowledged.
5. Decide how intervenability will be measured, and label it as a local measure.

Where teams usually fall short

• Safety reported as a count of defects with a safety label, which measures nothing about harm.
• A hazard warning that is displayed but never acknowledged, so its effectiveness is unknown.

What happens next

Operational constraint and risk identification set the frame. Fail safe, hazard warning and safe integration are the three with the most direct evidence.