Comment by user_7832 | Hacker News Reader

Comment by user_7832 | Hacker News Reader

Parent Full thread

(Caution, my thoughts are a bit rambly and scattered. It's been a while since I studied this in detail for a class where I had deep divided into the max 8 crash and FAA's lapse, so my memory is a bit foggy.)

The short answer is that unfortunately it's practically almost impossible to do this any other way, short of massively increasing funding for the FAA (which is presumably politically not going to be done.)

The long answer is, anyone with expertise ends up going to work at an aerospace company where they are properly compensated. You end up with the engineers at Boeing et al, and the regulatory folk at FAA etc. (This isn't unique to just aviation btw but most industries where you need a high level of technical expertise.) And if you think it should be easily doable... here's a thought experiment. Say 20 people in boeing know about MCAS (pre crashes and their publicity). If even only 2 FAA employees need to be technically sound in their knowledge of such a niche system, that's about 10% of how many boeing employees know about it.

Now if you extrapolate that to all systems and all employees across multiple companies... you see why it's an issue for one agencies to have such deep knowledge about everything. (And this isn't even considering what happens when people quit etc.)

Because of this, the most obvious answer to "how does the FAA, without much engineers, decide what's safe or not", is to ask the experts and just have them certify stuff, well, not under oath, but as close to that as possible.

>You end up with the engineers at Boeing et al, and the regulatory folk at FAA etc.

Incorrect. The engineers acting as FAA delegates are paid by Boeing, not the FAA. Toxic incentive. Stop shining that apple Leroy, plane needs to get shipped, and bossman is getting testy.

As far as your arguments on the information propagation front go, MCAS was deliberately designed against industry best practices and perception managed from regulators to do everything possible to avoid simulator training. Stop assuming good faith. This was malice. Malice that could only be executed upon because Boeing deliberately appointed junior engineers who didn't know the right questions to ask to FAA delegate positions.

I will not buy that delegating to the company is fine. The only people who benefit from that are the companies. I don't give a damn about the convenience to someone incentivized to go and industrially produce the least safe air transport product they can get away with. I care about the public's safety given I already know the market selects for incentive setters that flirt with risk in dangerous ways. And yes, that means we, the public, should damn well pay for that regulatory edifice. Especially if the rest of the world will treat FAA certification with reciprocity.

You are talking about type certification which is a different certification/process way prior to this. This is talking about 'certificate of airworthiness' for a specific tail number. This is confirmation that the aircraft/tail number was built to type design. It has nothing to do with defining/certifying type design.

The FAA cert of airworthiness just confirms that this tail number was build exactly to type design, or, if required, anything that didn't went through non-conformance (say someone accidentally drilled a hole in the wrong spot and engineers had to decide if that compromised type design, mitigate it if needed, document it, and include documentation on all that with the tail number). There is no malice here this position/certification doesn't do what you seem to think it does.

You salary concern is wrong too. Aviation is a small world. The QA person has to be approved by the FAA to have a job. Boeing can threaten a paycheck, the FAA a career. Boeing is also Union in Seattle.

But again you should not be asking questions at this level just confirming the tail number was manufactured exactly to type design. You should be doing type design checks, that were approved as part of the type design process, and signed off and approved by the FAA. If the system requires someone to be an aerospace engineer level asking smart question during this phase, the entire FAA system failed and it wouldn't matter if FAA was on site or not. This is 1-2 filing cabinets worth of documentation that checks were performed and everything conformed to type. Documenting/verifying every part's serialized number in case of grounding/recall. That assemblies were assembled in the type design specified order, not the faster way Bob on the floor came up with (I think this was the problem with the doors, contractors not doing to type design and type design not requiring auditing in a way that caught that). Stuff like that.

I'm afraid you're mostly preaching to the choir. My comment was to show how the real world issues develop ("How can the FAA trust Boeing to self audit after all that?!?" - because there's no other true option without major org changes) and how it's more complex than what meets the eye, not trying to justify status quo. I'm a very much pro safety guy who wants to do much more for aviation safety than just spread awareness on HN.

Btw for context, back in the 90s, a senior FAA official (I think a Director) had said something along the lines of “The FAA does not and cannot check everything, we just see that companies are doing their tests.”

> >You end up with the engineers at Boeing et al, and the regulatory folk at FAA etc.

> Incorrect. The engineers acting as FAA delegates are paid by Boeing, not the FAA. Toxic incentive. Stop shining that apple Leroy, plane needs to get shipped, and bossman is getting testy.

I wasn't talking about who gets paid by whom. I meant whether the FAA is putting out openings for regulatory/engineering people, and what kind of people Boeing is hiring.

> MCAS was deliberately designed against industry best practices and perception managed from regulators to do everything possible to avoid simulator training. Stop assuming good faith.

I never said Boeing was in good faith. My report (based off the excellent NTSB/DOT report significantly, if anyone wants to delve in depth) was pretty much about that.

> I will not buy that delegating to the company is fine. The only people who benefit from that are the companies. I don't give a damn about the convenience to someone incentivized to go and industrially produce the least safe air transport product they can get away with. I care about the public's safety given I already know the market selects for incentive setters that flirt with risk in dangerous ways. And yes, that means we, the public, should damn well pay for that regulatory edifice. Especially if the rest of the world will treat FAA certification with reciprocity.

I fully agree. Got a better solution/alternative in mind that's feasible?

(Disclaimer, I'm not from the US if that is of any importance.)

You seem to want a system where the regulators, while embedded at Boeing, are just wandering around and looking at things. What we have in aviation is a system where engineers, the people wearing the rings, with an ethical and moral obligation to prevent harm to other people, are charged with identifying issues and escalating them as far as the FAA if needed as they work on their day-to-day tasks.

The latter is a much better system to find issues than the former, you get a chance to see a lot more issues when your actively working on the systems where the issues would be than if your walking around and looking at paperwork.

Would the same system work for other industries, e.g. banks or medical research?

Reply on news.ycombinator.com