the original conception of golomb is far more intuitive than others, particularly when taking into consideration
this guy does a great job of reasoning much like Turing and co reasoned in 1943, using ratios in a hyperbolic computation graph space that reasons with correlations (contrasting null points on the distinguished circle – like enigma rotor points – with points on the overlying triangle – which represent non-unitary correlations linking plaintext to ciphertext).
now it becomes very obvious why Turing, in On permutations, is so adamant to set the mean of vector length to be 1. He is entirely reasoning in proportions.
It will be fun to see if Wildberger can get us all the way from here – at as he says the elementary stuff that is key to “thinking differet” – to fourier transforms computed in proportional algebras.
We know folks in the cryptanalytical attack on Tunny made exactly that leap.