3 ms·

What kind of problem do you consider it to be?

by sdinsn 7y ago

mehrdadn 7y ago
https://moneymaven.io/mishtalk/economics/boeing-737-max-major-design-flaws-not-a-software-failure-rVjJZBVzZkuZLkDJn3Jy8A/ https://moneymaven.io/mishtalk/economics/boeing-737-max-majo...
- Someone1234 7y ago
  Great link. I strongly agree that calling it a "software problem" undercuts the major engineering defects the aircraft has. The 737 Max doesn't even have enough physical sensors to supply the software with enough data to conduct true voting logic (something typically found on "safety critical" systems). The spec also didn't even have the software use both sensors it did have to detect defective inputs. The aircraft design was faulty before even one line of code was written. This wasn't a software bug. It wasn't a software design defect. The actual specifications and the physical design of the aircraft itself were flawed, and software was poorly used to plug the gaps.
phire 7y ago
It's not a software issue because the software was working excatly as specified. There is no bug. And its not really an aerodynamic issue either. Boeing have a massive procedural issue where they didn't do proper safety analysis of how the system works as a whole. The fact that they didn't identify this failure case as a major issue makes you wonder what other failure cases (on other planes too) that they missed. The original version of MCAS only moved the stablizer by 0.6 degrees, and only in situations with high angles of attack AND abnormally high g loading. That's the version they ran the safety analysis on, it relies on two types of sensors that shouldn't fail simultaneously. Later they modified MCAS to move the stablizer by much larger amounts (2.6 degrees, 4x larger) and they removed the high g loading restriction so it would operate in more situations. This also means it was now relying on a second sensor. And they never re-ran the safety analysis.
- magduf 7y ago
  Exactly. The whole thing is a gigantic failure of systems engineering, and if they made this big an error there, what error similar errors are lurking in their other planes? Then, to make matters worse, their response to the problem showed that they simply cannot be trusted.

What kind of problem do you consider it to be?

by sdinsn 7y ago

What kind of problem do you consider it to be?

https://moneymaven.io/mishtalk/economics/boeing-737-max-major-design-flaws-not-a-software-failure-rVjJZBVzZkuZLkDJn3Jy8A/ https://moneymaven.io/mishtalk/economics/boeing-737-max-majo...

Someone1234 7y ago
Great link. I strongly agree that calling it a "software problem" undercuts the major engineering defects the aircraft has. The 737 Max doesn't even have enough physical sensors to supply the software with enough data to conduct true voting logic (something typically found on "safety critical" systems). The spec also didn't even have the software use both sensors it did have to detect defective inputs. The aircraft design was faulty before even one line of code was written. This wasn't a software bug. It wasn't a software design defect. The actual specifications and the physical design of the aircraft itself were flawed, and software was poorly used to plug the gaps.

phire 7y ago

It's not a software issue because the software was working excatly as specified. There is no bug. And its not really an aerodynamic issue either. Boeing have a massive procedural issue where they didn't do proper safety analysis of how the system works as a whole. The fact that they didn't identify this failure case as a major issue makes you wonder what other failure cases (on other planes too) that they missed. The original version of MCAS only moved the stablizer by 0.6 degrees, and only in situations with high angles of attack AND abnormally high g loading. That's the version they ran the safety analysis on, it relies on two types of sensors that shouldn't fail simultaneously. Later they modified MCAS to move the stablizer by much larger amounts (2.6 degrees, 4x larger) and they removed the high g loading restriction so it would operate in more situations. This also means it was now relying on a second sensor. And they never re-ran the safety analysis.

magduf 7y ago
Exactly. The whole thing is a gigantic failure of systems engineering, and if they made this big an error there, what error similar errors are lurking in their other planes? Then, to make matters worse, their response to the problem showed that they simply cannot be trusted.