I’m not necessarily arguing his case, but if their software was open source it could be scrutinized far more closely. Is this software actually audited at all before an accident occurs? Do regulatory agencies have the bandwidth to do that? I really wouldn’t be surprised if this (and a lot of the embedded software in cars as well), simply isn’t being thought through because of commercial pressures.
The whole "enough eyeballs make all bugs shallow" thing is an empirical falsehood as evidenced by all of the breathtaking security bugs discovered in open source software over the past decade or so. It's not just anyone who is qualified to do auditing. If the software needs to be audited more carefully, the solution is to find the people qualified to do so and have them do it. It is almost completely unrelated to whether or not the source is open, which I'm also not against and in fact would be all for. However, arguing that proprietary code is inherently dangerous and open source code inherently safe is incredibly dangerous of itself and needs to be shut down.
You're confused about open source software and security.
If a business is using proprietary software that is found to have critical security bugs, they are 100% beholden to the vendor to get a fix. The vendor may decide to not fix that bug.
With free software, when (not if) a bug is found, the user has the right to fix the software.
While 'open source' might not be inherently 'more safe' than proprietary code (which we can't audit, so unlikely to be safe), it's inherently more safe due to the fact that the end user might actually be able to fix it, if required.
I didn't say that and the poster above didn't say that, but having the source available would undoubtedly help. In this case in particular you would have had many groups looking at the code in parallel, rather than the need for a slower, closed process because of commercial sensitivity. If both crashes end up being linked to this system, then the fact the code wasn't publicly available would likely have contributed to the inability to prevent the second crash.
You can’t say that though. For one thing code reviews are almost hopeless for catching normal bugs in the first place. For another, just because code is available does not mean anyone is actually looking at it. The people who are qualified to do this stuff quite likely will not do it for free.
This is about investigating one particular widget which has been implicated in these crashes, had the source been available people would have been able to look at exactly how that system works over the last months, rather than speculating while they wait for the closed investigation to occur, and while they wait for the second accident to occur. This isn't a general 'bugs will be found', it's about being able to substantiate targeted suspicions about a particular part of the code.
