Jon Skeet reports today (source) that :
Math.Max(1f, float.NaN) == NaN new[] { 1f, float.NaN }.Max() == 1f
Why?
Edit: same issue with double also!