Hacker News Reader

Comment by jjav | Hacker News Reader

> If you're wondering - hey, that sounds terrible, why doesn't the FAA do anything about it? Answer, because it practically can't without maybe a 10x or more budget boost that's practically impossible.

You're correct in that the FAA can't (or at least, it would be wildly impractical) have a staff of in-house experts that equal or exceed the experience of the manufacturer engineers on every possible topic.

But, that's not necessary to enforce a process of quality. It can be done by enforcing that the company must establish and document the processes to ensure quality, including which experts need to sign off on what. Then the FAA needs to hold the engineers who sign off professionally responsible and their managment legally liable if any circumvention is discovered down the line.

View on HN

> It can be done by enforcing that the company must establish and document the processes to ensure quality

You have no ideea.

It's sooo extremely easy to fake everything! I'm working with ISO15189 right now. I could just fill paperwork saying I did things that I didn't really do. Nobody could tell the difference. Sometimes there's an audit. It's announced months in advance and they only just look at the paperwork. I'm the only one here that knows the analysers have trace logs and how do get/print them. Some of those logs are just CSV text files. I could fake those too.

According to ISO, the company should do an audit every year. Last one was 3 years ago, but I'm certain that someone wrote down every year that they did it.

falsified paperwork is better than no paper trail in the event planes start falling out of the sky - spotting one moderate falsification could very well lead to further verification when it comes to investigations deemed important by the public.

Ah, yes, no paperwork, I completely forgot about that. It's even more common than fake doing stuff.

Don't be so confident that falsification can be spotted. It usually can't. Or require very time-consuming cross-checks that I never saw happen in my entire 20y career, such as comparing the number times an internal control was run vs. the quantity of consumables used in the process that were bought.

In case of Boeing, the other problem is... who's going to look before the plane crashes?

But when the plane finally crashes, wouldn't it be obvious in the investigation? Wouldn't that lead to further inspections of other planes?

But the procedure/QC already failed - those people are dead...

Yeah, but sometimes "if your negligence kills these people we will put you in jail" is exactly how you get people to pay attention. The vital part is closing the loop.

I disagree. That's not how it works. Responsibility is a product of education and environment, not a product of fear. Fear makes people avoid responsibility, not assume it.

Fear is a phenomenal backstop to responsibility, and in fact most aviation safety relies on the fact that pilots don't want to die.

Paperwork should be regularly audited and discrepancies investigated. The other nice thing about fear is that it should be very specific - people should know exactly what they have to do to not feel it.

> It's sooo extremely easy to fake everything!

The consequences part that I mentioned is where it's at, if one wanted to design a functioning system.

If there's an incident and the investigation reveals an engineer signed off on something that was contrary to process, they lose their license and get to switch careers to something else.

If it turns out management pushed the engineer to do that, personal liability on management means they get to do some long jail time.

With those two enforcement points, you cut out nearly all the cheating.

Personal liability on management never happens. If it did, they would find a way to transfer it somewhere else: a contractor from 3rd world country, a discardable subsidiary, etc. Nothing will stand in the way of profits, not even bad publicity - they would just sell / be bought, rebrand, and do it all over again.

> Personal liability on management never happens.

Oh, of course it'll never happen. I'm just responding to the thread that if a government actually wanted to create an effective quality mechanism, it's possible.

And I'm telling you that it can't. Well, maybe except for governement owned enterprises...

>But, that's not necessary to enforce a process of quality. It can be done by enforcing that the company must establish and document the processes to ensure quality, including which experts need to sign off on what. Then the FAA needs to hold the engineers who sign off professionally responsible and their managment legally liable if any circumvention is discovered down the line.

Having seen how quality processes work, it's really not a substitute for competent oversight. You can have a crap design and get all the paperwork is completely 100% in order, in fact the paperwork then makes it a lot more time consuming and expensive to fix the design!

^ this. My father is an aerospace engineer. He has worked for many companies and in various positions, director/head of engineering/certification across various sectors. He now works for the regulator as an auditor and specialises in the sectors he has worked in and often with companies he's worked for. There is a huge amount of integrity involved for the profession and I've never heard of any cases where this has been questioned. At the end of the day, there are rules and processes to follow, don't follow them and it's quite simple you loose the power to verify your own work, no business wants this and it's much easier to work with the regulator and conform to the rules. Also if the regulator determines there is an issue or fault then they can ground the affected aircraft, in a number of cases the airlines/manufacturer even have clauses with these companies that are the along the lines of "if your product grounds us, you will pay us x amount per day" The regulator will also specify how much cost to rectify work will be so the company will lose money, that's why it's easier to just follow the rules!

I should add that the regulator will be to the letter. I was told a case where the company followed the test form a spec from standard x, but standard x was superseded twice. The latest spec and standard x, the testing was exactly the same but because the company had done the testing against x and not the latest they had to redo all the work. Regulator would not shift or give dispensation on passing it.

Reply on news.ycombinator.com