3 releases (breaking)

0.10.0 Jun 28, 2022
0.9.0 Mar 22, 2022
0.8.0 Mar 6, 2022

#2088 in Algorithms

MIT license

70KB
2K SLoC

Allowance

Math representation of the physically needed permissible deviation of measures in rust avoiding floating point inaccuracy.

Based of type Measure with a accuracy of 1/10th my-meter (= 0.1μ).

Exaxmple

use allowance::AllowanceValue;

fn main() {
    let width1 = AllowanceValue::new(100.0, 0.05, -0.2);
    let width2 = AllowanceValue::with_sym(50.0, 0.05);

    // Adding two `AllowancesValue`s is strait-forth.
    assert_eq!(width1 + width2, AllowanceValue::new(150.0, 0.1, -0.25));

    // `!` inverts the direction of tolerance to /subtract/ measures.
    assert_eq!(!width1, AllowanceValue::new(-100.0, 0.2, -0.05));

    // Adding an inverted `AllowanceValue` wides the tolerance.
    assert_eq!(width1 + !width1, AllowanceValue::new(0.0, 0.25, -0.25));
}

Limits

Allowance

A 128bit wide value. Based on a Measure (64bit) for the value it could handle sizes up to +/-922_337_203 km with a variance of +/-214 m (Measure32).

Allowance64

based on a Measure32 (32bit) for the value it could handle sizes up to +/-214 m with a variance of +/-3 mm (Measure16).

No runtime deps