Sure, 0.1 can be represented exactly. But that's about the only issue with floating-point that that solves.