Reaction: Computer "integers" pretty much are actually integers - just with a limited range. Floats, however, are
NOT real numbers - they are poisoned from Square Zero by their limited precision.
Any and every time that you use floats, you have to make extremely sure that their piss-poor precision won't turn your program's results into sh*t.
Also - if your org can't be bothered to run long-duration tests of your life-critical real-time control software, your org's managers need to be fired and probably blacklisted.