3 ms·

100% of those software engineers have a plethora of safety checks that are applied continuously and repeatedly before their work is ever used in an environment

100% of those software engineers have a plethora of safety checks that are applied continuously and repeatedly before their work is ever used in an environment where any harm might actually be done to anyone. Someone operating a backhoe for example might dig into a gas main, swing their boom around and crush or decapitate someone, swing it across power lines and electrocute themselves, etc. Yes, mission critical software can kill people if it's not done right. That never happens because the developer was tired of a morning. So, back at ya with that honesty bit.

d0lph 9y ago
I guess as long as those catch 100% of the issues. Even still you're solving the problem on the backend instead of the source.
ajnin 9y ago
> 100% of those software engineers have a plethora of safety checks Well, this is simply not true. I suspect you conjured that figure out of thin air. Case in point, the famous Therac-25 accidents caused by a software bug causing large over-irradiations of cancer patients : https://en.wikipedia.org/wiki/Therac-25 https://en.wikipedia.org/wiki/Therac-25
- kungtotte 9y ago
  Yes, software from 40 years ago is absolutely relevant to this discussion. You think there are a lot of developers using those same practices today? Bugs happen, but with software you have several layers of safety checks to prevent them from becoming dangerous or fatal. A common one is having the code reviewed by one or more other developers. The odds of all those developers being unable to detect bugs at 9 AM are kind of long, wouldn't you say?
  - bartread 8y ago
    Not really, sad to say. I had a PR approved during the day yesterday by one of the smartest guys I've ever worked with and that passed all tests yet, due to a subtle issue, caused a problem in production this morning. It wasn't the end of any worlds, but this kind of thing happens all the time with complex systems.

d0lph 9y ago

I guess as long as those catch 100% of the issues. Even still you're solving the problem on the backend instead of the source.

ajnin 9y ago

> 100% of those software engineers have a plethora of safety checks Well, this is simply not true. I suspect you conjured that figure out of thin air. Case in point, the famous Therac-25 accidents caused by a software bug causing large over-irradiations of cancer patients : https://en.wikipedia.org/wiki/Therac-25 https://en.wikipedia.org/wiki/Therac-25

kungtotte 9y ago
Yes, software from 40 years ago is absolutely relevant to this discussion. You think there are a lot of developers using those same practices today? Bugs happen, but with software you have several layers of safety checks to prevent them from becoming dangerous or fatal. A common one is having the code reviewed by one or more other developers. The odds of all those developers being unable to detect bugs at 9 AM are kind of long, wouldn't you say?
- bartread 8y ago
  Not really, sad to say. I had a PR approved during the day yesterday by one of the smartest guys I've ever worked with and that passed all tests yet, due to a subtle issue, caused a problem in production this morning. It wasn't the end of any worlds, but this kind of thing happens all the time with complex systems.