Answer from cs61c-bg (Spencer Ahrens 16187781) for Question 3
Some older machines lacked guard digits in double precision floating point opperations, leading to rounding errors and incorrect results when calculating things like 1.0 - x