pub const NUMERICAL_TOLERANCE: f64 = 1e-9;
General numerical tolerance for floating point comparisons