This one hits deep. One of my first more complicated projects was a driver for a device that counted a lot of values. Imagine an energy meter.

At the beginning I thought about writing code to create my own type since I was limited to uint32 as the biggest on board type. To spare some work I quickly decided to use a float type for the first time without thinking much about it (I usually use fixed point integers for everyday work)

Way into the project I thought about what happens, when the actual value is gigantic and I try to add a very small number. After thinking about it I realized that nothing happens. So my counter would stop working. I changed everything and never used a floating point number again in my life I 🤣