Another attempt at describing the operation of the code wheel
Hm. Instead of trying describe the appearance of the full version of the thing, I’ll try describe how to construct a simplified version that works the same way.
Draw a circle. The code wheel is designed to be printed out on standard US printer paper. That’s eight and a half by eleven inches, so draw an 8 inch circle.
Draw a slightly smaller concentric circle inside it. Call the space between the two the rim of the circle. In the full code wheel we want to divide the rim into 26 equally spaced segments, but that’s a pain to describe so let’s divide our example into four segments instead. Just draw a vertical and then a horizontal line through the center of the concentric circles.
Now you have an eight inch circle with a thin rim that’s divided into four equal segments. And a big blank space in the center that’s also divided into four equal segments.
Go clockwise around the rim and write letters in the rim segments: A, B, C, D.
Set that piece of paper aside and get another. Draw a slightly smaller circle, so that the outer diameter of the new circle is roughly the same size as the inner diameter of the rim of the big circle. Repeat the rest of the steps, so we end up with two circles that look the same except one is slightly smaller than the other.
Cut them out. Punch a hole in the center of each of them, and stick them together with a paper brad, so that we can rotate them. With the A on the big circle lined up with the B on the little circle all the other letters also line up. Rotate the big circle a quarter turn clockwise and now the big A lines up with the little B, big B with little C, and so on. Another quarter turn and big A lines up with little C, big B with little D, and so on. In other words each configuration of our two circles implements a different four letter Caesar (or shift) cipher.
But that’s kinda boring. So we take the circles apart again.
On the big circle we write a letter in each of the empty segments in the center. Let’s be boring and just copy the letters from the rim. So now each quadrant of the circle has a letter on the rim and the same letter in the middle of the previously-blank space.
Now that we’ve added more letters to the big circle, we need some way to see one of them when the little circle is on top of it.
The little circle has a thin rim and a big empty center and they’re both divided into four quadrants. We want to cut out the empty center part of one quadrant and just leave the rim.
In practice we also have to leave a hub in the center of the little circle, for the hole the brad goes through, but I don’t think we need to worry about that here.
Anyway, with a hole in one quadrant, when the wheel is assembled we can see a letter on the big circle through it, and the letter displayed changes as circles rotate. Because we didn’t cut the little circle’s rim the operation of the two rims is exactly the same.
Does that all make sense so far?
The difference between what we’ve got here and the “real” code wheel is basically just the number of letters in the center of the big circle.
Take the circle apart again. Call the letters we wrote in the center segments a ring. We’re going to replace the single ring we had before with two rings. So first we erase the ring we have.
We lightly pencil in two concentric circles smaller than the rim, and along each of them we write letters the same way we did when constructing the single-ring version earlier. So now we have three sets of letters on the big circle: one along the rim, and two in the previously-empty center area.
If we reassemble the wheel then the hole in the little circle will show two letters: both of the rings in a single quadrant of the big wheel.
So we retrieve the wedge we cut out of the little circle and cut it in half so that one half covers the smaller ring and the other half covers the larger ring. Does that make sense?
We then tape the piece covering the bigger ring in place, so now the hole only shows the letter on the smaller ring.
Finally we cut another hole out of the quadrant that’s clockwise from the one that already has a hole. Instead of cutting out a big wedge like we did the first time, this time we only cut out enough to show the larger ring of the big circle. That is, exactly the same sized piece as the one we just taped back in place over the original hole. We could even use that piece as a template.
After cutting the second hole, then we have two windows. One of them shows the smaller ring in one quadrant, the other shows the larger ring in the next quadrant clockwise.
That’s pretty much it.
With the “real” code wheel there are 26 radial segments instead of four, there are six text rings instead of two, and instead of two windows you make six in the same way (one window for each ring, the windows starting at the smallest ring and moving one radial step clockwise and one ring out until all six rings have windows). And after making six windows in that way you make six more in the same way on the opposite side of the circle, for a total of twelve rings in two sets.
On accessibility
In terms of making the feelie accessible, part of the problem is that which bits are important and how they’re supposed to relate to each other is part of the puzzle.
As I mentioned before, in the “real” code wheel the rims have “words” in a strange script instead of letters. For our purposes let’s imagine that the big cirlce’s rim has words that start with the letter F, written in red: “Fee”, “Fi”, “Foe”, “Fum”, and so on. The little circle’s rim has words that start with the letter B, written in black: “Bee”, “Bi”, “Boe”, “Bum”, and so on.
In game, let’s say there’s a grandfather clock and on its face you notice the word “Foo” written in red at 12 o’clock and the word “Bar” written in black at 3 o’clock. So that would suggest that you want to arrange the code wheel such that “Bar” on the inner rim is 90 degrees clockwise of “Foo” on the outer rim. The text displayed in the windows then might be used for a password somewhere else.
That’s a made-up example (with twenty-six radial segments none of them are at exactly right angles to each other), but that kind of thing needs to be possible. That is, it isn’t just lining up one A with the other A and that kind of thing.