e, nor with very small numbers in layout. Adapted from
     * https://floating-point-gui.de/errors/comparison/.
     *
     * @param float $a
     * @param float $b
     *
     * @return bool
     */
    public static function lengthEqual(float $a, float $b): bool
    {
        // The epsilon results in a precision of at least:
        // * 7 decimal digits at around 1
        // * 4 decimal digits at around 1000 (around the size of common paper formats)
        // * 2 decimal digits at around 100,000 (100,000pt ~ 35.28m)
        static $epsilon = 1e-8;
        static $almostZero = 1e-12;

        $diff = abs($a - $b);

        if ($a === $b || $diff < $almostZero) {
            return true;
        }

        return $diff < $epsilon * max(abs($a), abs($b));
    }

    /**
     * Check `$a < $b`, accounting for inaccuracies in float computation.
     */
    public static function lengthLess(float $a, float $b): bool
    {
        return $a < $b && !self::lengthEqual($a, $b);
    }

    /**
     * Check `$a <= $b`, accounting for inaccuracies in float computation.
     */
    public static function lengthLessOrEqual(float $a, float $b): bool
    {
        return $a <= $b || self::lengthEqual($a, $b);
    }

    /**
     * Check `$a > $b`, accounting for inaccuracies in float computation.
     */
    public static function lengthGreater(float $a, float $b): bool
    {
        return $a > $b && !self::lengthEqual($a, $b);
    }

    /**
     * Check `$a >= $b`, accounting for inaccuracies in float computation.
     */
    public static function lengthGreaterOrEqual(float $a, float $b): bool
    {
        return $a >= $b || self::lengthEqual($a, $b);
    }
}