Toyota Case: Single Bit Flip That Killed

eetimes.com

1 pointkrupan2 comments

I'm not sure how to react to this. On the one hand it makes me think, shame on Toyota for having such poor code quality. On the other hand I ask myself, would any codebase survive 18 months of scrutiny by a team of industry experts that only need to prove that a bug in the code could have caused a malfunction? As a consumer the first really worries me (and more particularly as a Toyota owner, though I'm almost more worried about the products I use who's code hasn't yet been scrutinized by Michael Barr). As an engineer the second worries me.

Concern about that second case is an important thing to have as an engineer, don't get me wrong, we should definitely put effort and care into making products safe and reliable. One thing, however, that was missing from the analysis of the Toyota code (at least from what I read) was the estimated probability of a code malfunction occurring and causing a problem.

Nothing is perfect and simply pointing out potential failure modes without characterizing further their likelihood under the various conditions that could cause them seems like plain old fear mongering (which, I suppose, served the purposes of the plaintiff in this case precisely).