I saw a bunch of suggestions involving ridiculously overcomplicated solutions, but they all looked like they would take me more than an hour to code up, so they were quickly discarded. I took the route of doing everything in binary arithmetic, and through a curious quirk, only division supports floating-point answers (for example: "1/2" produces "0.5" as instructed, but if you turn around and try to multiply that by "1", you get "0"). Actually, it's not really that curious a quirk. I got lazy in the middle and just started casting everything to integers after seeing that none of the test