4 ms·

The difference is that airplane software standards are by far the most rigorous in the world and so rigorous that most software developers are shocked at the le

by Veserv 5y ago

The difference is that airplane software standards are by far the most rigorous in the world and so rigorous that most software developers are shocked at the level of rigor employed where as driverless cars have exactly zero software standards with respect to their safety or fitness for purpose. Comparing the processes in play is a outrageously fundamental category error; the systems have standards that are like 6 orders of magnitude apart. As a example, just look at the 737 MAX, a airplane viewed as a literal deathtrap and a complete indictment of airplane software development, which had 2 crashes in 400,000 flights. At a average flight distance of ~500 miles, the 737 MAX, a airplane literally multiple orders of magnitude worse than any other and a complete failure of the approval process, only resulted in 1 fatality per 100 million person-miles. That is safer than the driving average of 1.1 fatalities per 100 million person-miles. What is viewed as by far the most atrocious outcome in the airplane sector, a system literally hundreds or thousands of times more dangerous than the average, is above average in the car sector; that is how far the gap is. And this is ignoring driverless car software which is being developed to even lower standards than regular car software and which has no actual independent quality requirements. In contrast, far more lives are at risk with respect to driverless cars, so they should be developed to standards probably around 1000x better than airplane software which puts them around 0.0001% of the way there to a acceptable level of quality even if the software was already as reliable as regular car software.

jklinger410 5y ago
> The difference is that airplane software standards are by far the most rigorous in the world and so rigorous that most software developers are shocked at the level of rigor employed... You mean the standards set by the FAA, which regularly defers to the private industry to regulate itself, due to complexity that it can't afford to keep up with? Those standards?
- olyjohn 5y ago
  Flying is still the safest way to travel. So I'm not sure what you're complaining about with the FAA.
  - nitwit005 5y ago
    The crashes revealed the FAA wasn't doing the job people thought they were. We achieved safety in this case by grounding all of the planes after crashes, which is hardly the ideal method.
- mostertoaster 5y ago
  Yeah this is how most standards go. I use to work on medical devices, and the safety requirements for them were developed by my company and the other big players in the industry, and those became the FDA standards. The FDA just wants there to be standards, and the companies have a vested interest in making sure those standards ensure safe medical devices.
- lcnPylGDnU4H9OF 5y ago
  I'm lost by your comment. It seems that you're trying to suggest either 1) that the standards as mentioned necessarily aren't True Standards - as a blanket statement - because they're set by the industry or 2) that these standards aren't True Standards - specific to this particular discussion - because there is no reason to think they're necessarily more rigorous than the standards for driverless car software. If you're trying to suggest something else, please understand that it is unclear. I'm all for a dissenting opinion, but when it is not presented in an understandable way the discussion becomes hard to follow. (I don't mean to suggest that the argument is wrong when I call out the True Scotsman vibes; just that it is the argument I see.)
  - jklinger410 5y ago
    What is the legal/regulatory difference between the aerospace industry regulating itself vs the auto industry regulating itself in regards to self driving? Besides the obvious that the aerospace industry has a head start.
    - lcnPylGDnU4H9OF 5y ago
      I believe I understand your point better. From the person you're responding to: > Yes, those standards that produced a safety regime of 20 years of flights without software being even implicated in a single fatality. ... > As you note, unfortunately the standards as employed in practice have recently regressed dramatically to allow a system as unsafe as the 737 MAX to fly. ... Your point, if I'm not mistaken, being that the FAA has clearly allowed for less stringent regulations in recent years. (Please correct me if I'm wrong. You can ask questions but it does not seem that they lead so immediately to the answers you expect.) That being said, please consider that I am genuinely trying to understand you: > I'm all for a dissenting opinion, but when it is not presented in an understandable way the discussion becomes hard to follow.
- Veserv 5y ago
  Yes, those standards that produced a safety regime of 20 years of flights without software being even implicated in a single fatality. Literally trillions if not quadrillions of passenger-miles per software-related fatality. Those same standards which result in a system like the 737 MAX rightfully being excoriated for only being 10% safer than driving and resulted in successful demands for grounding of the entire fleet. As you note, unfortunately the standards as employed in practice have recently regressed dramatically to allow a system as unsafe as the 737 MAX to fly. So it seems only prudent to go back to the historical, far more rigorous, process which did result in the amazing historical safety regime. And, since self driving cars are a even harder and more dangerous technology with far more downside risk, they should use a standard even more rigorous than the historical airplane standards which are still by far better, from a safety perspective, than any other consistently applied software process, such as cars, by multiple orders of magnitude.
  - jklinger410 5y ago
    > Literally trillions if not quadrillions of passenger-miles per software-related fatality. So these were all miles where people were unwilling testers as well. The difference being airplane auto-pilot is primarily focused on keeping the plane in the air, not dodging pedestrians. I think it does a lot of warnings for the pilots as well, but I'm not in the industry. It's really comparing apples to oranges, isn't it? > since self driving cars are a even harder and more dangerous technology with far more downside risk, they should use a standard even more rigorous than the historical airplane standards So this is on The United States Department of Transportation to be far more involved in regulating the software that governs self driving. But the question is, was the FAA as relaxed in governing the aerospace industry 60 some odd years ago as USDT is now, and did they just get lucky that flying planes is actually easier than driving cars? Or is there actually some negligence involved?
bsder 5y ago
> The difference is that airplane software standards are by far the most rigorous in the world and so rigorous that most software developers are shocked at the level of rigor employed Now they are. This regulations are a result of several tragedies culminating in a big disaster: https://en.wikipedia.org/wiki/American_Airlines_Flight_191 https://en.wikipedia.org/wiki/American_Airlines_Flight_191

by Veserv 5y ago

jklinger410 5y ago

> The difference is that airplane software standards are by far the most rigorous in the world and so rigorous that most software developers are shocked at the level of rigor employed... You mean the standards set by the FAA, which regularly defers to the private industry to regulate itself, due to complexity that it can't afford to keep up with? Those standards?

olyjohn 5y ago
Flying is still the safest way to travel. So I'm not sure what you're complaining about with the FAA.
- nitwit005 5y ago
  The crashes revealed the FAA wasn't doing the job people thought they were. We achieved safety in this case by grounding all of the planes after crashes, which is hardly the ideal method.
mostertoaster 5y ago
Yeah this is how most standards go. I use to work on medical devices, and the safety requirements for them were developed by my company and the other big players in the industry, and those became the FDA standards. The FDA just wants there to be standards, and the companies have a vested interest in making sure those standards ensure safe medical devices.
lcnPylGDnU4H9OF 5y ago
I'm lost by your comment. It seems that you're trying to suggest either 1) that the standards as mentioned necessarily aren't True Standards - as a blanket statement - because they're set by the industry or 2) that these standards aren't True Standards - specific to this particular discussion - because there is no reason to think they're necessarily more rigorous than the standards for driverless car software. If you're trying to suggest something else, please understand that it is unclear. I'm all for a dissenting opinion, but when it is not presented in an understandable way the discussion becomes hard to follow. (I don't mean to suggest that the argument is wrong when I call out the True Scotsman vibes; just that it is the argument I see.)
- jklinger410 5y ago
  What is the legal/regulatory difference between the aerospace industry regulating itself vs the auto industry regulating itself in regards to self driving? Besides the obvious that the aerospace industry has a head start.
  - lcnPylGDnU4H9OF 5y ago
    I believe I understand your point better. From the person you're responding to: > Yes, those standards that produced a safety regime of 20 years of flights without software being even implicated in a single fatality. ... > As you note, unfortunately the standards as employed in practice have recently regressed dramatically to allow a system as unsafe as the 737 MAX to fly. ... Your point, if I'm not mistaken, being that the FAA has clearly allowed for less stringent regulations in recent years. (Please correct me if I'm wrong. You can ask questions but it does not seem that they lead so immediately to the answers you expect.) That being said, please consider that I am genuinely trying to understand you: > I'm all for a dissenting opinion, but when it is not presented in an understandable way the discussion becomes hard to follow.
Veserv 5y ago
Yes, those standards that produced a safety regime of 20 years of flights without software being even implicated in a single fatality. Literally trillions if not quadrillions of passenger-miles per software-related fatality. Those same standards which result in a system like the 737 MAX rightfully being excoriated for only being 10% safer than driving and resulted in successful demands for grounding of the entire fleet. As you note, unfortunately the standards as employed in practice have recently regressed dramatically to allow a system as unsafe as the 737 MAX to fly. So it seems only prudent to go back to the historical, far more rigorous, process which did result in the amazing historical safety regime. And, since self driving cars are a even harder and more dangerous technology with far more downside risk, they should use a standard even more rigorous than the historical airplane standards which are still by far better, from a safety perspective, than any other consistently applied software process, such as cars, by multiple orders of magnitude.
- jklinger410 5y ago
  > Literally trillions if not quadrillions of passenger-miles per software-related fatality. So these were all miles where people were unwilling testers as well. The difference being airplane auto-pilot is primarily focused on keeping the plane in the air, not dodging pedestrians. I think it does a lot of warnings for the pilots as well, but I'm not in the industry. It's really comparing apples to oranges, isn't it? > since self driving cars are a even harder and more dangerous technology with far more downside risk, they should use a standard even more rigorous than the historical airplane standards So this is on The United States Department of Transportation to be far more involved in regulating the software that governs self driving. But the question is, was the FAA as relaxed in governing the aerospace industry 60 some odd years ago as USDT is now, and did they just get lucky that flying planes is actually easier than driving cars? Or is there actually some negligence involved?

bsder 5y ago

> The difference is that airplane software standards are by far the most rigorous in the world and so rigorous that most software developers are shocked at the level of rigor employed Now they are. This regulations are a result of several tragedies culminating in a big disaster: https://en.wikipedia.org/wiki/American_Airlines_Flight_191 https://en.wikipedia.org/wiki/American_Airlines_Flight_191