Ender left a good comment on my post about Somebody Else's Risk. There's an interesting counter-argument to my posting: you could say that since I believe myself a good programmer, I SHOULD take on these life-risk programs, because I will do them better than the other guy. At the time, I instinctively mistrusted this argument. After the many years that have passed since then, I think I know why. It's the nature of computer products that every good one is duplicated many times. People rush to fill every perceived niche to overflowing, whether it's an operating system, blogging software, or an email system. Each time I decide to take the risk of writing software where people's lives are at risk, I am ADDING one more approach to all the others. I'm not doing it instead of someone else, and there's little risk that the software I contribute to will be so good that it chases all the inferior ones away. And the bugs I leave in my software may affect some unlucky person.
Writing software, or doing any work, where the quality affects people's lives, is simply a grave responsibility that has to be taken seriously.