Hacker News Reader

Comment by sho_hn | Hacker News Reader

These are the most comparable times in the history of Airbus:

https://admiralcloudberg.medium.com/one-hundred-seconds-of-c...

https://admiralcloudberg.medium.com/thinking-like-a-computer...

Unlike with Boeing they didn't feature intentional obfuscation and fraud, but very similar themes of the airplane's software model of what's going on diverging from the pilots' and resulting in disaster.

Along with stories like the Therac-25, this is one of my "favorite" engineering stories relevant to my profession.

View on HN

From the article:

> Although it got off to a rocky start, the A320 went on to achieve a better safety record than most traditional aircraft types. And although there have been a couple of close calls, no Airbus has ever crashed because of the sort of computer failure that skeptics so deeply feared.

Airbus didn't just fight the need to improve their designs. They just kept improving on their designs and fixed what needed to be fixed.

I'm not sure why Boeing and Airbus are being treated equivalently by you and the other commenter. Especially since the two situations are nothing alike. Boeing hid MCAS to avoid a new type rating. The Airbus A320 just had new types of systems that would continue to be iterated on over time and were simply new.

> I'm not sure why Boeing and Airbus are being treated equivalently by you and the other commenter.

I think you misunderstood my intention in posting these links. I'm actually not sure how, since I pointed out clearly in the original text how the cases are very different. I posted them with the intent of "this is arguably the most like that, which is still very different". Maybe to hammer it down more, I gauge the Boeing case to be criminal, the Airbus examples not, and it's worth comparing the manufacturer conduct in these cases.

The common theme is software assists increasing complexity and the likelihood of the operator's thinking to diverge from the machine. I find that interesting personally and professionally (I make safety-critical vehicles with high degrees of automation for a living). If you want some interesting reading, I still recommend them.

> very similar themes

implies that Airbus and Boeing should be treated similarly. When the issue isn't necessary the automation or software, it's how each company behaves in situations where there are serious faults or deficiencies with their designs.

> implies that Airbus and Boeing should be treated similarly.

Honestly, I think that's you misreading the language.

> how each company behaves in situations where there are serious faults or deficiencies with their designs

Which the comparison can illustrate. Rather than downvoting me and somehow trying to hide information, why don't you just double-down on what I did and extend the argument? I see exactly no reason for conflict here. What I originally wrote (this is misconduct, this is not) seems to exactly agree with you.

In any case, I suppose I have more of a creator's view and am more interested in what makes a good software assist vs. a bad one, and for me any assist implicated or involved in an incident are interesting data points in the spectrum.

Which is what Boeing used to do.

Wasn’t the original 737 the plane that had a number of early accidents and people literally didn’t want to fly on it? Maybe I’m thinking of a different model.

But they fixed it and treated it properly. And it became one of the most popular planes ever. Everyone trusts them.

Then they made the MAX version and made obviously stupid decisions like not having redundancy on the MCAS sensors. Things that the history of aviation tells you will go wrong. And they lied about it and covered it up, and lied about and covered up and obstructed the investigation into their plants not finishing planes correctly. And…

Boeing had a great reputation for a reason.

Reply on news.ycombinator.com