fn one_minus_square(y: f64) -> (f64, f64)
1 - y² as an unevaluated sum high + low, accurate far beyond a double.
1 - y²
high + low