That, right there, lines up exactly with what I was afraid had been going on. I'm quite impressed with Brazil for doing their own footwork to ensure their pilots are well informed.
I am abjectly horrified at the self-certification process the FAA has adopted, however.
As someone who has spent a long time doing Quality Assurance, I've learned that Engineering often develops blind spots as a result of diplomatic friction with other areas of a company, and cannot often fend off potentially dangerous technical compromises without an adversarial force capable of stopping the entire show. Even with one, if that department is still answerable to the same overarching business pressures, there is still an elevated level of slippage and inherent risk.
The FAA should never have allowed self-certification. By acting as an external quality control force, the FAA guarantees a decoupling of any financial interest on the part of the inspector, and actually acts as a unifying force against which competing airframe manufacturers can unite. It aligns incentives to keep the FAA getting better at design inspection/approval, and gets airframe manufacturers focused on making information relevant to certification as easily accessible as possible.
Trying to implement that type of mechanism in house, or through delegation aligns incentives in such a way where poor or inaccessible documentation is encouraged to make any inquiries or attempts to prove liability more difficult. It also increases the incentive for the in-house inspector to "give-in" through social forces, and workplace dynamics.
I recognize these types of dynamic everyday in software shops, but I'd never imagined it would happen somewhere that produced something that absolutely must work.
2 comments
[ 3.2 ms ] story [ 18.5 ms ] threadShould be a bipartisan effort to put public interest representatives into the FAA.
I am abjectly horrified at the self-certification process the FAA has adopted, however.
As someone who has spent a long time doing Quality Assurance, I've learned that Engineering often develops blind spots as a result of diplomatic friction with other areas of a company, and cannot often fend off potentially dangerous technical compromises without an adversarial force capable of stopping the entire show. Even with one, if that department is still answerable to the same overarching business pressures, there is still an elevated level of slippage and inherent risk.
The FAA should never have allowed self-certification. By acting as an external quality control force, the FAA guarantees a decoupling of any financial interest on the part of the inspector, and actually acts as a unifying force against which competing airframe manufacturers can unite. It aligns incentives to keep the FAA getting better at design inspection/approval, and gets airframe manufacturers focused on making information relevant to certification as easily accessible as possible.
Trying to implement that type of mechanism in house, or through delegation aligns incentives in such a way where poor or inaccessible documentation is encouraged to make any inquiries or attempts to prove liability more difficult. It also increases the incentive for the in-house inspector to "give-in" through social forces, and workplace dynamics.
I recognize these types of dynamic everyday in software shops, but I'd never imagined it would happen somewhere that produced something that absolutely must work.