[Blog Post] Cracking a medieval code

[MarcoP]

With respect to other substitution ciphers, this method is interesting because it is "lossy": the text is actually degraded since, for instance, the symbol 'f' in the encrypted text can read 'f' or 'e'. This makes the cipher somehow similar to an abjad: information is lost, but someone knowing the original language has no difficulty reconstructing the original.

BTW, this cipher reduces both first order (character) and second order (digraph) entropy: the effect on conditional entropy (the difference between second order and first order entropy) is minimal.