Ogre,
You are correct, I only had the UK in mind (although I once had cause to read DEF(AUST) 5679!). You say "...to prove we have done everything reasonably practicable, with requests for details of what mitigations we didn't implement and why not." Who is asking this? The legal authorities or the procuring agencies? There was/is a similar issue in the UK. The legal authorities only come into play when, as in this sad case, something goes wrong. The procuring agencies seemed to be the ones asking such questions despite the legal duties only applying to the manufacturer. It can be argued that accepting the Def Stan (00-56) in the contract allows this investigation but it doesn't alter the law of the land nor those charged with responsibility for its observation.
While the UK Health And Safety At Work Act uses "reasonably practicable" it doesn't attempt to define it as it represents a relatively well understood UK legal concept. When the Act was published in 1974, it's contents weren't new, the Act just codified what the Common Law already said about health and safety. I still believe that the translation of SFAIRP into something meaningful for engineers isn't very straightforward. The HSE is quite good at this - their web pages are quite informative - but the guidance has limited utility for manufacturing complex products like military aircraft.
I agree that there can be hard side effects on manufacturing companies, but mostly those whose products have the highest severity hazards in use. I believe such companies need to recognise this and use a safety engineering process from product concept to delivery and in-service. Companies need to recognise that safety engineering is a fundamental part of the Systems Engineer's toolbox. For me Def Stan 00-56 (in any of its early versions) hasn't really been very good at making this clear and in my experience this helps lead PTs to address safety just before entry into service. This approach puts the emphasis on process outputs rather than continuous monitoring and involvement which is the way if they're truly interested in "details of mitigations". I often thought that the requirements of earlier versions of Mil Std 882 tended to be better about the use of safety through product life. Funnily enough, the Eurofighter contractual process is based on the Mil Std as the Def Stan hadn't been published when the contract was signed.
EAP