This is a repeating decimal:
Take it apart
0.352 + 0.000171717 = 0.352171717
as e^(i*pi) showed in his post:
you can see that 352/1000 can be reduced
so add the fractions
need a common denominator
and that can be reduced:
divide numerator & denominator by 625 to get the fraction in lowest terms.
The continued fraction algorithm is the method of choice for doing this type of conversions.
The above is not that algorithm.
Perhaps someone will explain/demonstrate the continued fraction algorithm for converting decimals to rationals.