Parsing the provided numbers to two IEEE 754 double-precision floats, and then subtracting them, yields 2.0 (due to rounding) In particular, 9999999999999999.0

Parsing the provided numbers to two IEEE 754 double-precision floats, and then subtracting them, yields 2.0 (due to rounding) In particular, 9999999999999999.0 rounds up to 10000000000000000.0 (note, this behavior depends on the rounding mode, https://en.wikipedia.org/wiki/IEEE_754#Rounding_rules https://en.wikipedia.org/wiki/IEEE_754#Rounding_rules )