The company failed 33 of 89 audits during an examination conducted by the Federal Aviation Administration after a panel blew off an Alaska Airlines jet in January.
It’s a measurement on an arbitrary scale. Nothing I’ve said is news to anyone who designs safety critical systems. I’m certainly not saying that safety isn’t important or that we can’t assess it. What I’m saying is that placing a number on that assessment will always stray into the realm of politics in a way that physics and mathematics never does. It lulls ignorant people into the belief that something is safe or not safe. They feel safe because they’ve been told it is safe or vice versa. Physics doesn’t care if you feel safe.
It’s notable that contemporary safety standards such as ISO 26262 generally avoid numerical assessments, for the reasons outlined above.
First Incidents per hour is not arbitrary. These numbers compare very well to daily activities such as walking, driving, bathing, eating, swimming so that non specialists have a good idea of how much risk an activity carries by comparing it to an activity they’re familiar with.
Secondly ISO 26262 produces ASILs as its output which are qualitative, but still based on probably assessments in terms of chance of incidence per hour. The reason for qualitative instead of quantitative assessments of the more general SILs (based on IEC61508, the parent of ISO 26262) is that qualitative is cheaper than quantitative and the automotive industry is full of corner cutting.
Third, aircraft use QUANTITATIVE risk assessments based on ARP476, so risk can be directly measured and mathematicaly compared to any other activity. When people say “flying is safer than driving” it’s not arbitrary, it’s based on real math. The same math the FAA is using to find safety issues in the Boeing production line.
Fourth
I’m certainly not saying that safety isn’t important or that we can’t assess it.
It’s a measurement on an arbitrary scale. Nothing I’ve said is news to anyone who designs safety critical systems. I’m certainly not saying that safety isn’t important or that we can’t assess it. What I’m saying is that placing a number on that assessment will always stray into the realm of politics in a way that physics and mathematics never does. It lulls ignorant people into the belief that something is safe or not safe. They feel safe because they’ve been told it is safe or vice versa. Physics doesn’t care if you feel safe.
It’s notable that contemporary safety standards such as ISO 26262 generally avoid numerical assessments, for the reasons outlined above.
First Incidents per hour is not arbitrary. These numbers compare very well to daily activities such as walking, driving, bathing, eating, swimming so that non specialists have a good idea of how much risk an activity carries by comparing it to an activity they’re familiar with.
Secondly ISO 26262 produces ASILs as its output which are qualitative, but still based on probably assessments in terms of chance of incidence per hour. The reason for qualitative instead of quantitative assessments of the more general SILs (based on IEC61508, the parent of ISO 26262) is that qualitative is cheaper than quantitative and the automotive industry is full of corner cutting.
Third, aircraft use QUANTITATIVE risk assessments based on ARP476, so risk can be directly measured and mathematicaly compared to any other activity. When people say “flying is safer than driving” it’s not arbitrary, it’s based on real math. The same math the FAA is using to find safety issues in the Boeing production line.
Fourth
Is this you?