We have worked on two problems from cryptography, as tests of our methods for identifying errors in expert reasoning. Gordon has also created two cipher texts, using different principles from the usual crytography approaches.
The Voynich Manuscript
This is a book several centuries old, whose text has never been deciphered. It is richly illustrated, with a mixture of strange and ordinary-looking images. Previous researchers had assumed that the regularities in its text were too complex to have been meaningless gibberish; Gordon found a way of producing these regularities using a low tech method of generating meaningless gibberish text. The implication is that the manuscript is probably a hoax. The image below shows part of a page from the manuscript.
The D’Agapeyeff Cipher
This is a short challenge in a 1939 book, which was an introduction to code making and code breaking. It has never been deciphered. Gordon and Gavin have published a series of blog articles about this. There are several claimed solutions by different researchers; the solutions are all different. The image below shows the cipher.
Gordon’s cipher texts
Gordon produced two cipher texts while working on the Voynich Manuscript. One was similar in appearance to the Voynich Manuscript, and gave some useful insights into the practical issues involved in creating an illustrated cipher text. The other was deliberately very different, to investigate what happens when you deliberately depart from the usual assumptions in code making. Neither has ever been cracked.
The Ricardus Manuscript: Detail from one page
The Penitentia Manuscript: One of the pages