> If you want to pretend that quality matters, move to aerospace, finance, or any critical systems.
The important word here is pretend. If the system is safety critical you'd think that quality matters. It doesn't. What matters is following the process laid out in ISO 26262 or whatever standard applies to your field. Whether or not that improves quality is not important (it does, but only a little). The important part is that you have reduced/no liability for accidents if you can prove that you did everything by the book.
Yep, part of the quality standard for medical devices is to have processes, and also processes to improve and change the processes. So theoretically there shouldn't be any doubt as what to do in a given circumstance when something fails, and that standards would rise as development matures. But in reality it's just a bunch of paperwork, and as far as the software is concerned, standard practices of unit and integration testing, code reviews, and strict compiler settings are more than adequate already.
Edit: It's all about documenting that those things have been done and signed off on, so keeping track of paperwork.
...where he talks about his role and questions regarding the Challenger shuttle disaster; I don't know if things have changed much since then at NASA, but I suspect they haven't, given Columbia and such.
Mmmyes, but the process implies a heavy chain of verifications. So even when there has been an idiot and a lazy guy, at some later stage in the process, it has to go through someone who cares at some point and it gets fixed at that point.
After all, that is the point of those heavy processes: manage to product something fairly reliable despite the acknowledged presence of morons and shitty managers. The process does not try to eradicate those people, it works around them with some kind of redundancy.
Having been one of those 'concerned' people, I must admit it is exhausting. You know that if some work comes from this guy, or that company, it will be 90% shit and you have to explain them again and again what is wrong and how to fix it, or do it yourself when you are really fed up. So basically you do your job + the slackers' job. That is how the system works. As a system, the result is not too bad. As far as individuals are concerned, however...
The important word here is pretend. If the system is safety critical you'd think that quality matters. It doesn't. What matters is following the process laid out in ISO 26262 or whatever standard applies to your field. Whether or not that improves quality is not important (it does, but only a little). The important part is that you have reduced/no liability for accidents if you can prove that you did everything by the book.