I played with creating a logarithmic slider thing [1] in Javascript that I hoped I could package up as a kind of "widget" people could use on their web pages. But I don't really know Javascript that well—or rather how to make an API out of a Javascript thing.
Anyway, to test it I tried to make an Ohm's Law calculator [2].
I would love to see a site like the one in this post have some kind of interactive slide rule on the web page itself.
[1] https://github.com/EngineersNeedArt/SlideRule
[2] https://www.engineersneedart.com/ohmslaw/index.html (the yellow slider is not directly user-moveable in this example)
Robert A. Heinlein, Have Space Suit—Will Travel (1958)
I still have the slide rules, so this post was a great rabbit hole to go down. In software there's no need for them but I still find them fascinating as a window into how engineers used to get their work done.
... but in the Real World they work pretty well for the sort of calculations you might need to do in the field (literally, in a field, sometimes) and don't require batteries, are reasonably waterproof, and reasonably robust if dropped.
They're pretty useful for teaching amateur people how to implement algorithms. Multiple ways to solve problems, some easier than others, some more efficient than others, with immediate rewards of faster higher accuracy.
Never thought of that, and I used to work in an ATEX environment where calculators powered by watch batteries had to be carefully logged and carried across to a "safe" area inside a special (horribly expensive) Peli case.
* 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)
* https://www.youtube.com/watch?v=oYQdKbQ-sgM
"Professor Herning" (?) also has a good series of videos on the use of various scales as well:
* https://www.youtube.com/@ProfessorHerning/videos
His playlist starting at the beginning (C and D scales) with a Manheim layout:
* https://www.youtube.com/playlist?list=PL_qcL_RF-ZyvWJJkJOk_O...
* https://sliderulemuseum.com/Manuals/M37_Post_Manheim_Instruc...
Some manuals / books on slide rules:
* 1909: https://archive.org/details/mannheimsliderul00coxwrich
* 1922: https://archive.org/details/cu31924002978561/mode/2up
I never spent the time to get quick with it, but I could absolutely see it being quicker than a calculator. You’d just have to be aware of the limits to its precision if you were in a field that required it.
One problem with a slide rule is that it only performs operations on normalized mantissas. You have to keep a parallel exponent calculation in your head, and that slows you down. Also, maintaining best precision slows you down.
For multiplication, the DLDP in the result is:
- the sum of the DLDPs of the multiplicands MINUS 1 if the multiplication is done with the slide sticking out to the right of the ruler's body (for example 2.0 x 3.0 = 6.0).
- the sum of the DLDPs of the multiplicands if the multiplication is done with the slide sticking out to the left of the ruler's body (for example 5.0 x 4.0 = 20.0).
There's a similar rule for division, but that's left as an exercise for the student.
We were taught to estimate and use the rule to refine. I date back to the early electronic calculator era and we still had textbooks referencing slide rules etc.
"I want a dropping resistor for a plain old 1980s LED in a car" (back in ye old red LED 20 mA days) "Well experience indicates that will be far more than 500 ohms and somewhat less than 1K and IRL you're probably going to install a 680 and call it good" If you want an actual calculation for engineering purposes you calculate the ideal value under worst case conditions as about 585-ish ohms or whatever using the slide rule, purchasing LOLs at the idea of buying 0.1% precision resistors for mere LEDs, installs cheap 680 ohms and ships it. Maybe 680s if you want it bright to see in daylight or 820 if you want better odds to survive an alternator field winding dump or open battery (about the same thing). You can at least use the slide rule to verify everyone rounded in the "safer" direction to handle the worst case scenario.
I did keep a slide rule as a backup for exams in college when calculators were still LED but never really used one after a couple of years in high school.
They do what people want, the keyboard feel is infinitely smoother than tapping on a phone, etc.
Yeah, the 12C was the standard in business school. But I needed a new calculator and the 41 with its various modules worked fine and was more general purpose.
We have speed electronic calculators now instead of slide rules, but they give a wronger answer and people aren't even aware of it or know why.
This article: “lol, is that the depth of your commitment”