3 ms·
Deep scrutiny only has moderate impact when it comes to finding and fixing bugs. An issue can be extremely difficult to identify and understand in the first pl
by user5994461 7y ago
Deep scrutiny only has moderate impact when it comes to finding and fixing bugs.
An issue can be extremely difficult to identify and understand in the first place. Assuming you can find it anyway, the fix is very costly both in financial and political terms. All things considered, most problems just remain under the carpet.
For a comparable case. Think of the unwanted acceleration in the Toyota car. A software bug that caused multiple deaths. Toyota didn't admit to having any bug. The first audit found glaring quality issues but no explanation for the unwanted acceleration. It's only few years later that another audit identified how to trigger it, well, 10 different ways to make it happen.
- cglace 7y agoReading through the Wikipedia page on the Toyota pedal issue it claims it was a physical problem.
- user5994461 7y agoLast article and discussion on this: https://news.ycombinator.com/item?id=9643204 https://news.ycombinator.com/item?id=9643204 If you read the long text and the attached documents, there are multiple software, hardware and design issues detailed. Some deadly ones around acceleration about tasks scheduling and priority inversion, well known aspects of embedded real time software with known solutions to avoid them. The lead is well buried. That's another obstacle with non-trivial debugging. Nobody has the attention span, the time and the knowledge to understand non-trivial bugs, let alone when they're intentionally kept in the dark and pressured by companies.