TIL: Float.MIN_VALUE and Double.MIN_VALUE are not the most negative values

Float.MIN_VALUE and Double.MIN_VALUE are the smallest positive values, not the most negative. The most negative value is -Double.MAX_VALUE.

The name MIN_VALUE means different things for whole numbers and decimals:

• Whole numbers: Int.MIN_VALUE is the most negative value, -2,147,483,648.
• Decimals: Double.MIN_VALUE is the tiny positive number closest to zero, 4.9E-324 (0.000…0049 with 323 zeros after the decimal point). It is not negative.

So for Double, the “min” is about how close to zero it can get, not how low it can go. This matters when you search for a maximum:

var best = Double.MIN_VALUE        // Bug: this is 4.9E-324, a positive number
for (x in doubleArrayOf(-5.0, -2.0)) {
    best = maxOf(best, x)
}
println(best)                      // 4.9E-324, not -2.0

Every negative input is smaller than this starting value, so it never gets replaced. The fix is to start from -Double.MAX_VALUE (the most negative finite Double) or Double.NEGATIVE_INFINITY. Int.MIN_VALUE doesn’t have this problem, because for Int it really is the lowest value.