It seems plausible enough to me. Many comp sci undergrads would be dimly aware that floating point arithmetic is notoriously difficult to get right and can often lead to surprising errors if you get it slightly wrong, and also dumb enough to believe that e^π^−π is exactly 20.