There's a lot to be said about programmers, but in how many other professions do people actually feel like crying for not finding the best solution to a problem?
I used to work on simulations of industrial lifting cranes. Crane operators would use the simulator to plan a lift before going out into the field to execute it. This was 'round about 2008, when New York City had all of those cranes fall over in one year. Luckily, none of them were any of our customers, but every day was a new test of patience as we waited on the news to tell us what company was responsible so we could check out customer list.[1]
I'm now working on DOT road signing systems. The software I am writing will be used to figure out when and where to put up speed limit and warning signs.
What is really, super concerning is that, though I am capable of doing all of the math and physics that the job requires (and there is really no reason for you to believe me on that point, which makes the situation even worse), none of the other programmers I've ever worked with on these projects could.[2] The companies that hired me just got lucky that I typically go with the least-effort job application process and am too lazy to be constantly applying to jobs. There was no vetting process, it was just "take the cheapest guy we can get".
When people could die as a direct result of a programming error on your behalf, it tends to change your perspective on things.
[1] When the company ignored my analysis that certain parts of the data was logically inconsistent (this system used look-up tables to do everything, very little was actually simulated), I quit. As far as I was concerned, the fact that none of our own customers had been in the news was mere luck and I had no interest in being on watch with my hands tied behind my back when it finally did happen.
[2] The most recent fun defect I had to fix involved someone having added two angles together, one in degrees, the other in radians. That prompted this little thought-experiment in code: https://github.com/capnmidnight/UnitsOfMeasure/tree/master/U.... It works for some simple things and has an interesting side effect of helping document the code, which I realized when I failed to convert a particular poorly documented section of the code to this system.
You should. We don't certify programmers, so it's a very real risk.
On the other hand, I do not support the efforts of my state to attempt to license software engineers. As it's written, it really only seems like a protectionism racket to keep upstarts from competing with established firms. To whit: employees of S-corps are exempt. Way to miss the entire freaking point, Virginia.
