Comments by "angeldude101" (@angeldude101) on "The Hidden Geometry of Error-Free Communication" video.

  1. Hm... how can these 24-bit/3-byte codes be expressed as a text string... Base-64 is a pretty good text encoding for binary data, so how many base-64 digits would be needed to encode a single codeword...? 2^24 = 16 777 216, which is the number of possible Goley codewords, and 64^4 = 16 777 216... Well, that was easy. So it's possible to write any Goley codeword as 4 characters, each of which is an upper or lower case letter, a decimal digit, or one of 2 symbols. (Most implementations seem to use '+' and '/', but ',', '-', and '_' also seem to be used in some cases.) There's probably a more fitting error correction code for text. In general, I'm curious how possible it is to have a text-based error correction code such that someone can write down a short identitifier in messy handwriting and then have the code successfully correct for various ambiguities in reading said identifier to enter it in.
    1