Live data from Hacker News

Volkswagen's U.S. head: individuals engineered emissions cheating

reuters.com

141–150 of 259 posts

Re: Volkswagen's U.S. head: individuals engineered emissions cheating

#141
post #89
post #67

Earlier quoted context omitted.

Doesn't it blow your mind that there are no "standards of engineering" or whatever for software? There's no licensure body for software engineers who build software running your car, and therefore no accountability on a personal level. When an engineer builds a bridge, she has to personally sign off on the bridge, saying it's safe, and is risking not only her professional career, but I think she can also be jailed an…

There are many standards of engineering involving software. - https://en.wikipedia.org/wiki/ISO_26262 - https://en.wikipedia.org/wiki/IEC_62304 - https://en.wikipedia.org/wiki/MISRA_C

The big problem with all those standards, frankly, is that you've gotta pay money to actually evaluate them.

There aren't any guarantees that they'll be useful, that they'll match the modern development processes in your language, that they'll fit your problem domain, etc.

Those standards are there primarily to make the publisher a buck--not to represent the codified wisdom of up-to-date practitioners in a field.

Until we've got a truly open-source standard for people to code against, we should stop wringing our hands about these things.

Re: Volkswagen's U.S. head: individuals engineered emissions cheating

#142
It's an utterly brazen lie to say the company didn't know about this.

Anyone who's ever built anything as part of a team - whether the Linux kernel, a space shuttle, or a Lego set - can tell you design decisions like this cannot be secreted into the final product.

So let's say a software guy decides to make this fix. How do you do that?

- You'll have a version control system, won't you? At the very least someone will see it. There's some interesting if statements referencing the speed of the rear wheels, which you can explain at code review.

- How do you test it? I doubt your normal office happens to contain the entire running setup of a diesel engine, with nobody around to notice, and nobody needed to tell you about the details. Because you know, an engine test rig wouldn't have details, right? You just turn it on yourself, spin the car up to speed and debug, right? And I'm sure there's nothing complicated about gas measurement, even though it tends not to be a part of CS courses.

- What do you tell the cluelessly naive good guy engineer about his results? Would he really only test the car under test conditions? Does any engineer with even the slightest grain of curiosity test only the one mode that he's supposed to?

Re: Volkswagen's U.S. head: individuals engineered emissions cheating

#143
post #53

These are worrying times for VW if the new CEO to say something like this. Many VW engineers will be considering leaving if VW is going find a couple of engineers as scapegoats, and not be back them up.

Presumably there are a couple of engineers who deserve some of the blame, but wow, what better way to say "Our culture is poison". Both because they're throwing independent contributors under the bus, and because he's saying the culture is such that the engineers would think it's OK. To be fair, I think he might not have meant it to come out that way. EDIT: Looking again at the article, it seems like the US CEO isn't…

Unfortunately, we may never know the truth in the scenario. The way that automotive world works with the multi-tier supplier chain, anyone who is not close to the source, will never know where the real blame is.

Behind the scenes in VW's HQ in Wolfsburg, it must be chaos.

Re: Volkswagen's U.S. head: individuals engineered emissions cheating

#144

Nope, definitely systemic within the company. The engine is a system that must meet certain requirements. The standard way of meeting those requirements is to have a urea injection system. That's what the competitors in the US do; it's also what Subaru, for example, don't do, and they don't sell into the US market. In order to pull off this little 2-person hack that Horn alleges, those couple of software engineers wo…

Not only that but by finding themselves a scapegoat, management has ensured that in the future, honest mistakes will go completely unreported. These execs are clearly more interested in CYA than actual results.

Re: Volkswagen's U.S. head: individuals engineered emissions cheating

#146

Do we think VW are the only car manufacturer doing this? I very much doubt it. They're just the first to get caught. Where rules stand in the way of profit, they get bent or broken. Banking, automotive industry... which next? Oil? Housing? Pfff ... come on, we all know they're rotten, it's just a question of how rotten and whether it comes out. I suspect it's the tip of the iceberg for the auto industry. Whether we g…

everyone does this, so far I only saw only few independent tests, but in every one all diesels wre over limit

http://www.focus.de/auto/ratgeber/kosten/unregelmaessigkeite...

Re: Volkswagen's U.S. head: individuals engineered emissions cheating

#147
Uh, right. I don't think so.

I'm sure it went like this:

"Now, I'm not telling you to do anything illegal, but if you don't pass those emissions standards on the next deadline, you're fired."

At each level down the chain.

I actually saw this in action at a big company subcontracting division to two younger employees (one male and one female): "We need that contract. Period. Your expense account on this unlimited and unaudited. Make sure you go pull enough cash that you don't have to put anything embarrassing to the company on the card. I'm not telling you to do anything illegal, but if you don't close that contract, you don't have a job Monday. Anyhow, make sure the client has a good time this weekend. See you Monday. Maybe."

They closed the contract. No one asked what it took. I left that job post haste.

Re: Volkswagen's U.S. head: individuals engineered emissions cheating

#148

Nope, definitely systemic within the company. The engine is a system that must meet certain requirements. The standard way of meeting those requirements is to have a urea injection system. That's what the competitors in the US do; it's also what Subaru, for example, don't do, and they don't sell into the US market. In order to pull off this little 2-person hack that Horn alleges, those couple of software engineers wo…

You're making an awful lot of assumptions there.

>those couple of software engineers would have to have a major say over the entire engine design. I call bullshit.

The two people they're mainly talking about are head of R&D group and engine development. Yeah I'd imagine they have major say over engine design.

Re: Volkswagen's U.S. head: individuals engineered emissions cheating

#149

It's an utterly brazen lie to say the company didn't know about this. Anyone who's ever built anything as part of a team - whether the Linux kernel, a space shuttle, or a Lego set - can tell you design decisions like this cannot be secreted into the final product. So let's say a software guy decides to make this fix. How do you do that? - You'll have a version control system, won't you? At the very least someone will…

> What do you tell the cluelessly naive good guy engineer about his results? Would he really only test the car under test conditions? Does any engineer with even the slightest grain of curiosity test only the one mode that he's supposed to?

While I agree with the rest of your posts, I can't let this one slide. This happens ALL THE TIME. I have to beat engineers senseless over this on a regular basis.

"Just because it passed one test, one time, does not mean you are done. It has to do something sane even when the rest of the world is coming flaming down around it. Crashing could be the sane thing to do; logging could be the sane thing to do; returning very obviously insane values is non-ideal, but might be your only option. A couple of safety checks are warranted."

Re: Volkswagen's U.S. head: individuals engineered emissions cheating

#150
It's quite possible that the "company" didn't actually know about it because they built a culture where people know what when to break rules and when not to ask.

It's not all that different of a culture than a government agency that doesn't need to be told to do things that benefit the party of the current regime.

Post reply on HN