As I discussed in that post, using the BigDecimal (double) constructor rarely does what one would expect when passed a float or double. For a double, the Java developer is typically better off using ...