557 sats \ 3 replies \ @orthwyrm 5 Jan \ parent \ on: Stacker Saloon
WDYM? I do some hardware control in my day-to-day work.
At my previous company, I wrote so many tests for my code, I think it kind of became hardware, lol. I also treated this code as if I can never touch it again because it will be shipped out as hardware and thus must work in all circumstances, even if the sun explodes tomorrow.
And I always feel very suspicious if code works on first try. I have a strong urge to not trust my own code but rather green checkmarks that are generated by other code that I wrote (but is easier to verify)
Working at SN made me realize that just reading code can also be enough. You don't need to test everything just because you changed one line, lol. You just need to understand and make very good educated guesses, what might break and what will probably still work fine.
I still miss my 200+ green checkmarks though.
reply
Ahhh I see. Yeah, I understand.
It really depends on how critical the code is, I guess. Some of my stuff is simple data processing so it's not the end of the world if a calculation is off. On the other hand, some other code manages stage control for some verrrrrry expensive optics that can easily damaged if there's a stage-objective collision. Also we have lasers, so many things can go wrong with lasers...
I suppose working on SN must be similar. Simple UI stuff is low-risk, managing deposits and channels high risk.
reply
Simple UI stuff is low-risk, managing deposits and channels high risk.
Exactly.
At my previous company, I never had to think about this. All I did was to work on the same microservice that was connected to and consumed by basically everything else (microservice architecture at it's finest) and some bugs and their impacts were so ridiculous, I started to treat my code like any bug in it could grind the whole company to a halt, lol
And every time I found a bug, I wrote a test case for it so it never happens again.
Funny to think about it in hindsight.
reply