That's a dangerous line of reasoning in this case.
I could be convinced that this court ruling is erroneous, and that the unintended acceleration issues can be entirely accounted for by floor mats and driver error. But in this case I think we should be thankful for bad floor mats and driver error, as they've brought to light very fundamental flaws in Toyota's firmware engineering processes.
If there had never been a single unintended acceleration in a toyota vehicle it would not have been through robust engineering but instead through luck. And we need our vehicles to be safe by design, not through happenstance.
If you can solve the halting problem, we can make software that doesn't misbehave.
I'm not saying Toyota should be allowed to provide this shoddy piece of software in a critical subsystem, but I very much think a) other vendors software will be just as crappy and b) this feels like the court longing for reasons to fault Toyota on something that was still very likely user error, not software misbehaving.
It's not necessary to solve the halting problem to have the kind of protections aircraft have, let alone things like not drastically miscounting the stack space, having memory protection against stack overflow, using ECC RAM, not having 11,000 global variables, etc. Even if it was user error, this isn't even close to a sane design.
If there had never been a single unintended acceleration in a toyota vehicle it would not have been through robust engineering but instead through luck. And we need our vehicles to be safe by design, not through happenstance.
I especially appreciate your comment because my childhood best friend (an electrical engineer who designs safety-critical systems) thinks this way. He started out in avionics, and was the lead designer for the avionics system for a commercial airliner that so far has a very good safety record indeed, and then he moved over to the medical device industry. In his work, "zero defects" is the only standard, and fundamental understanding of how a system works, from the level of subatomic physics on up, is his approach to design with no hidden flaws. That approach is not easy, but he thinks that is the appropriate approach when human lives are at stake.
This problem in general is yet another symptom of the immaturity of the software industry as a whole. Most standards are community standards rather than well accepted and extremely well known official standards. And a lot of best practices still come down to judgment. Moreover, best practices and standards vary greatly depending on the nature of the product. Even within a small sub-field like embedded systems the requirements are very different for a car, airliner, or 3D printer.
It's telling that even at a big company like Toyota which is fairly risk averse and is well known for its commitment to quality and safety they are still capable of churning out pretty crappy software that is hugely important. Software dev is still a pretty hard problem overall, we live in an era where there have been lots of successes, but failure is still common and the consequences of failure can sometimes be severe.
I could be convinced that this court ruling is erroneous, and that the unintended acceleration issues can be entirely accounted for by floor mats and driver error. But in this case I think we should be thankful for bad floor mats and driver error, as they've brought to light very fundamental flaws in Toyota's firmware engineering processes.
If there had never been a single unintended acceleration in a toyota vehicle it would not have been through robust engineering but instead through luck. And we need our vehicles to be safe by design, not through happenstance.