Comment by buggi23

To add to this, there's a great set of resources by Joe Pasquale explaining the mathematical theory behind how various functions can be computed by slide rules:

* https://cseweb.ucsd.edu/~pasquale/Classes/SlideRule/

* Mathematical Foundations of the Slide Rule (PDF): https://cseweb.ucsd.edu/~pasquale/Papers/IM11.pdf

* Why Does A Slide Rule Work? (PDF): https://cseweb.ucsd.edu/~pasquale/SlideRuleTalkLasVegas14.pd...

The gist of it is:

1. First, define a way to represent any univariate monotonic function f(x) on a graduated scale. (Specifically: select a discrete set of x values, and for each of these x values, place a mark with label x at a distance proportional to (f(x) - f(x_L)) from the left endpoint, where x_L is the leftmost x value.)

2. Then, if we have two such scales f(x) and g(x) that can slide relative to each other, we can compute functions of the form h(x, y, z) = f_inverse(f(x) + g(y) - g(z)).

It ends up being surprisingly versatile -- the above resources show how you can compute:

1. Multiplication: x * y using f(x) = log(x) and g(y) = log(y), with z fixed at 1

2. Hypotenuse: sqrt(x^2 + y^2) using f(x) = x^2 and g(y) = y^2, with z fixed at 0

3. Parallel resistors: 1/(1/x + 1/y) using f(x) = 1/x and g(y) = 1/y, with z fixed at +infinity

4. Exponentiation: x^(y/z) using f(x) = log(log(x)) and g(y) = log(y)