this might be a hot take and i quite frankly havent thought about this too much but floating point values being the "default" way to handle any sort of fractional component even in tools designed to be accessible to non-insane people is a terrible idea, floating point values are good because they're a single format for a huge range but if you know the order of magnitude that your use case will need or even can vaguely approximate it you will almost always get more reliable behavior with fixed-point values. if i primarily expect fractional values between 0 and 100 why would I waste so many of my precious bits to poorly approximate 40 billion
tag yourself i'm cool
