I must be getting old:
On Oracle JDK 8:
class A {
public static void main(String[] argv) {
float f3 = Float.MIN_NORMAL;
float f3e1 = f3/2.0f;
System.out.println(f3 + " / 2 -> " + f3e1);
float f3e2 = f3/4.0f;
System.out.println(f3 + " / 4 -> " + f3e2);
}
}
Result
1.17549435E-38 / 2 -> 5.877472E-39
1.17549435E-38 / 4 -> 2.938736E-39
How is that possible? Isn't the Float.MIN_NORMAL the smallest Float still different from 0.0 (a Float in Java being required to be 32-bit IEEE 754 floating point, sadly without exceptions (WTF!), and thus the normal is ~1.17549E−38)?
In which case, the printed values must be Double or something else with more precision than Float ... but the values are declared as float.
