News for hackers. Stuff that matters.
Re:Volkswagen's U.S. head: individuals engineered emissions ... by bcantoni on Thursday October 08, 2015 @01:51PM (#10354535)
I'm assuming the development process for embedded auto software like this would be pretty rigorous, meaning it would be fairly difficult to sneak in code like this.

I also assume that testing of emissions software like this would involve a lot of lab/road testing with the vehicles against the EPA test profile and therefore a lot of people would be involved. Suggesting this is the stealthy work of just a couple of software engineers is very hard to believe.

Can anyone who's worked on embedded software in the automotive environment comment?

Re:Volkswagen's U.S. head: individuals engineered emissions ... by cnvogel on Thursday October 08, 2015 @02:13PM (#10354767)
The investigations regarding Toyotas(?) unexplained/unintended acceleration suggest otherwise. The code was apparently pretty bad, halfheartedly reviewed and just unmaintainable. While there the software seemed to have (a huge number of) errors put in by mistake, of course most independent reviews should also spot intentional mischief.

I don't claim that VW had comparably crappy code, but "industry practice" likely wouldn't have caught it either way.

Re:Volkswagen's U.S. head: individuals engineered emissions ... by joezydeco on Thursday October 08, 2015 @03:05PM (#10355240)
Re:Volkswagen's U.S. head: individuals engineered emissions ... by cnvogel on Thursday October 08, 2015 @03:13PM (#10355302)
Thanks for adding this!
Re:Volkswagen's U.S. head: individuals engineered emissions ... by zardo on Thursday October 08, 2015 @02:17PM (#10354803)
At a high level, I beleive VW uses the V-model. Everything implemented should tie back to a high level requirement, and it should have been verified.

https://en.m.wikipedia.org/wiki/V-Model