Post

I guess my thought is that substantially all the software I'm responsible for is "mission critical". I've always gravitated to problems that matter

3 reposts82 likes

I think a good judgement is "does someone die if you get it wrong". I have worked on Internet infrastructure and other consequential software, and even then I was surprised at the pack of rigour. But ultimately my boss at the time was like "yes it would suck if things went wrong, but who dies?"

True, but there's often a wide range of ways to be wrong, and when shipping, at some point it has to be "done". My boss's comment put it into perspective, and no one died. Since then I've been far more consequential environments and I keep coming back to that comment.

unbounded risk or not is separate from mission-critical in the way people mean when they draw this distinction, I think more like, how tight do your bounds need to be if you can never ship an update again (or if shipping an update takes an unbounded amount of time and effort)

Ease of updating is I would consider a different dimension, though most of the software I write is not SaaS so is harder to update than that

I could argue I work on mission-critical software and be right in a certain sense, but at the end of the day if that software has (in spite of the team's best efforts) defects, the nature of our software enables it to be patched in a matter of hours in an emergency

Yeah I think that like. The inventors of IPV4 couldn’t have possibly expected to fail this test in the future. Who puts IPv4 in medial equipment or airplanes?

That definitely describes some of the software I've written haha. "Oh this can't be hard.... oh wait the depth of complexity is endless. time to roll up my sleeves, the only way out is through"