# Does machine cryptography only consist of rotor machines?

For the chronological categorisation of cryptography, I have proceeded as follows:

1. Cryptography by hand (e.g. Caesar, Vigenére, etc., till around 1900/1920)
2. Cryptography using machines (e.g. Enigma, TypeX, etc., till around 1960)
3. Computational Cryptography (e.g. DES, AES, RSA, etc.)
4. The Future of Cryptography (e.g. Quantum crypto)

The slots 1,3, and 4 are well documented and have a lot of different systems. But I'm struggling with slot 2: The heart of all the systems I found were rotors. Some were operated by mechanics, some by electro magnetism, but all use rotors for their cryptographic applications.

My questions: Does cryptography using machines only consist of rotor machines like Enigma, Purple, Typex, etc.? If other design principles exist, can someone give me an overview of them and maybe examples?

• Do you count M209 as a rotor machine? M209 was easier to break than Enigma! Commented Feb 8, 2022 at 13:54
• I would say it is a portable mechanical rotor machine Commented Feb 8, 2022 at 14:07
• Don't forget the Jefferson disk (mixture of a mechanical device and manual encryption) en.wikipedia.org/wiki/Jefferson_disk Commented Feb 8, 2022 at 18:05

I would argue that a great many cryptographic designs lie between the mechanical designs of 2 and the micro-electronic designs of 3. These designs are electronic in nature, but without the universal computation capabilities afforded by microchips. Most of these designs were based around Linear Feedback Shift Registers (LFSRs), which are easy to instantiate with basic electronic components.

To quote section 16.3 from Schneier's "Applied Cryptography" (1996)

In the early days of electronics, they were very easy to build. A shift register is nothing more than an array of bit memories and the feedback sequence is just a series of XOR gates.

...

Most designs are secret; a majority of military encryption systems in use today are based on LFSRs.

Whether one classifies such electronic (vice micro-electronic) designs as machines, computational designs or a category to themselves is a question of terminology.

Examples might include:

In addition to the electronic but non-CPU machines mentioned by Daniel S, there are a variety of even simpler non-CPU cryptographic tools:

• Paper authentication systems such as DRYAD

• Some people distinguish between "electric rotor machines" that transpose the alphabet by sending electric current through moving cipher wheels, vs. purely mechanical "Pinwheel machines" ("Pin-and-lug cipher machines" ) which also have rotating wheels, but don't use any electric power.

• A variety of purely mechanical tools were used to speed up doing (effectively) complex scrambled-alphabet Beaufort Cipher (Vigenére Cipher): Le Sphinx, the Swiss FG cipher device, the Jefferson disk, etc.

• Tape cipher machines have very simple electronics, and speed up doing (effectively) very simple 1-bit XOR-based Beaufort Cipher (Vigenére Cipher) or one-time pads.

• Some of the earlier "secure speech / secure telephone" devices used analog scrambling electronics, some without any digital electronics.

• "grilles" and "turning grilles", mechanical tools to speed up transposition ciphers.

• message translation tables: Sprechtafeln and SLIDEX

• nomenclators and codebooks, which can sometimes be immune to even computer-aided cryptanalysis.

• The Solitaire cipher (designed around 1999 by Bruce Schneier) doesn't really fit into any of your categories.