Floats typically have a mantissa and exponent which has a very distinct bit pattern. When both are non-zero (or only the exponent) we likely have a float and should assume so and display the integer value only as a fallback (similar to how floats are displayed today).
Floats typically have a mantissa and exponent which has a very distinct bit pattern. When both are non-zero (or only the exponent) we likely have a float and should assume so and display the integer value only as a fallback (similar to how floats are displayed today).