Hacker News new | past | comments | ask | show | jobs | submit login
[dupe] Toyota Unintended Acceleration and the Big Bowl of “Spaghetti” Code (2013) (safetyresearch.net)
42 points by charlieegan3 on Nov 29, 2015 | hide | past | favorite | 29 comments




Is that why it just disappeared from the front page listing?

I can understand tweaking it down in the rating algorithm, but if there's enough people that want to chime in that it makes it to the front page even after it's been covered before, removing it entirely seems fairly heavy handed.


On HN, a story is considered a dupe if it has had significant attention in about the last year. We bury those, i.e. move them off the front page. If we didn't do that, the HN front page would quickly fill up with repeats, especially the most controversial repeats, which never run short of people wanting to "chime in".

The present case is especially clear cut because there were two major recent discussions of the story.


Fair enough, and I actually thought the top result in the search was this story the first time, so thought it was last discussed six months ago, not that and just a single month ago.


And yet business folk still probably think it's perfectly fine to outsource this kind of work because code is code, right?


Why do we never hear about programmers ethically challenging the auto makers when they clearly know about these flaws? It's akin to working on medical devices. Where is the incentive to not make lethal mistakes?


Part of it, I'm sad to say, is that many, many programmers don't see themselves as professionals, but rather as mere laborers. If you try to get a doctor or a lawyer or a structural engineer to do something egregiously wrong, they'll generally know that they're crossing a line. They see themselves as having a responsibility for the human effects of their work.

Many programmers, though, will do whatever they're told, even when they know it's pointless, wasteful, or dangerous. I've read literally hundreds of stories about programmers working on something that was egregiously fucked up. But I've heard damned few stories where programmers refused to do something they thought wrong and either a) got management to back down, or b) got fired for their refusal to do something unethical.

I don't blame individuals for that. We're not trained as professionals, we don't have a professional organization that most of us join, and we therefore don't have a serious professional code of ethics. It's unreasonable to expect most people to make ethical stands alone while creating a code on the fly. But I'd love to see the situation fixed.


It seems wise to have a programmers union. Maybe pg should put some efforts towards that instead of seed funding. After all, if we have a union, we can set standards for wages and quality. Certifications. And more. It'd to the world of professional programming some good.


To me unions solve an asymmetry where companies can fire a worker more easily than a worker can lose their job. So personally I'm not interested in a union of all programmers, as that would create a power asymmetry the other way.

I'd more thinking about a professional organization like the AMA, the assorted bar organizations, the society of professional journalists, or the various engineering societies. Basically I want an organization that can set forth reasonable professional ethics, do advocacy and education, and act as our collective voice to back up individual engineers who refuse to do something unethical.


These are not well educated programmers. They are probably run of the mill undereducated software engineers from an outsourcing company. They do not recognize the technical issue here, albeit report an ethical issue.

Companies like Toyota generally outsource software development to other companies like Bosch. Bosch employs a bunch of programmers in Bangalore who are just out of college, and pay them peanuts to develop a safety critical system.


  * they are part of a big enough team that individual responsibility is removed
  * they're not given enough time to produce good quality code
  * the design came from someone without sufficient expertise, and the programmers "just built it"
  * they did the best given the time they had
  * it passed testing
  * programmers are generally not empowered to make go/no-go decisions
  * even if they did express concern they may not have been listened to
  * any internal challenges on ethical grounds were probably kept quiet and worked around
Usually with this kind of incident there are multiple process and quality failures, not just "why did the programmers not do better"


I have always objected when software was designed against my own ethics. I was never forced out, I just got another job or worked on another project. Every developer asks the question "why" am I making this. Only developers who care about the product seem to dig when given a hand wave and not a lot of them seem to care.


Is it ethical to just walk away from a problem?


Is it a problem that people commission unethical things? If it was a danger to human life, I'd have to drop an anonymous note to a newspaper and regulatory body or two.


Would that be like a whistlerblower program?

Because we know how that always ends, with the person driven out of the company or set up to be fired.

The only way to get big corporations to behave is prison time for executives so they become paranoid about doing things right. Unfortunately in the USA we never ever seem to do that, we just give them a few million dollars in fines instead which they pay with the next month's profits (ala BP).


> Michael Barr...spent more than 20 months reviewing Toyota’s source code...in a hotel-sized room, supervised by security guards, who ensured that entrants brought no paper in or out, and wore no belts or watches.

Why the restriction on belts and watches?


Clearly you've never been a magician. ;) It's easy to hide things in the band with invisible slight of hand.


This is one reason why I will always prefer a manual transmission. Because it leaves me, the human, with a completely mechanical override. And it's more fun.


I suspect cars are more fly-by-wire than you realize. And wasn't this issue with the throttle control? I suppose you could have downshifted with either type of transmission, but most drivers aren't prepared for sudden mechanical failure like this.


Just kick on the clutch pedal and the power transmission is broken.


As long as the clutch linkage doesn't break :-)

This happened to my dad while waiting at for a light at an intersection. Transmission in first, foot down on clutch: suddenly car leaps into intersection. This was long ago on a Volkswagen bus.

The drive by wire which is terrifying is the power steering. Traditionally it's hydraulic, but more cars are using electric. I sure hope that an electrical component failure can not lead to full turning force.

http://www.caranddriver.com/features/electric-vs-hydraulic-s...


shutting off the car, setting it to neutral, etc... do the same thing. bad drivers will be bad.


> shutting off the car, setting it to neutral

Both of these are no fail-safe measures these days, as both ignition and auto transmissions are done entirely in software.


I'm not aware of any passenger car with an electronically controlled clutch. The driver in question finally resorted to pulling the handbrake, which locked up the wheels, so she lost control of the car. Flooring the clutch pedal OTOH, if she'd had one, would have saved her life.


Some people seem more likely to experience the problem than others.

http://www.thedetroitbureau.com/2012/04/unintended-accelerat...


IIRC it was generated C from matlab.


Imagine the recall for this.


Article dated 2013, not news.


From 2013




Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact

Search: