Relative Error

Version: Phuoc Hung OJ Extended

Problem Statement

Given an approximate value $x$ and a nonzero reference value $ref$, the relative error is $E=\dfrac{|x-ref|}{|ref|}$. Print the ratio itself, not a percentage, with exactly 6 digits after the decimal point.

Input

The only line contains the approximate real value $x$ and reference value $ref$.

Output

Print the relative error $E$ with exactly 6 decimal places and no % character.

Subtasks

• Subtask 1 (100 points): $x,ref$ are finite; $ref\ne0$.

Examples

Example 1

Input

99.9 100

Output

0.001000

Explanation. Substituting the input into the formula gives |99.9-(100)|/|100| ≈ 0.001. Formatting with six decimal places produces 0.001000.

Example 2

Input

1.01 1.0

Output

0.010000

Explanation. Substituting the input into the formula gives |1.01-(1.0)|/|1.0| ≈ 0.01. Formatting with six decimal places produces 0.010000.

Example 3

Input

-9.9 -10.0

Output

0.010000

Explanation. Substituting the input into the formula gives |-9.9-(-10.0)|/|-10.0| ≈ 0.01. Formatting with six decimal places produces 0.010000.