The complete register
Twenty-two systems, and how each one stopped working
Each entry states the promise the system made and the specific thing that broke it — usually procedure rather than mathematics.
Section 1 — Substitution
Section page →| Entry | Section | What it promised, and what broke it |
|---|---|---|
| One alphabet for another | Substitution | The oldest idea in the subject, and the one that letter frequency dismantles in an afternoon. |
| Frequency | Substitution | Language leaves a fingerprint that survives substitution intact. |
| Nomenclators | Substitution | A cipher alphabet plus a codebook of names and words, and the bulk that made it unwieldy. |
| Homophones | Substitution | Several symbols for one common letter, to flatten the fingerprint. |
Section 2 — Polyalphabetic
Section page →| Entry | Section | What it promised, and what broke it |
|---|---|---|
| The table that held for three centuries | Polyalphabetic | A grid of shifted alphabets with a repeating keyword defeated frequency analysis and was called indecipherable for three hundred years — until someone noticed that the keyword repeats, and that the repetition shows. |
| The keyword repeats | Polyalphabetic | The flaw that was there from the beginning and took centuries to be seen. |
| Running key | Polyalphabetic | Using a book instead of a word, and why that is better but still not enough. |
| Autokey | Polyalphabetic | Letting the message extend its own key. |
Section 3 — Machines
Section page →| Entry | Section | What it promised, and what broke it |
|---|---|---|
| Wiring it into metal | Machines | Rotors turn the substitution at every keystroke, so the alphabet never repeats within any practical message. |
| The stepping | Machines | How the wheels advance, and why the irregularity matters. |
| The reflector | Machines | A design decision that made the machine reciprocal — and gave away that no letter could ever encipher to itself. |
| Operator habit | Machines | Predictable openings, reused settings and lazy indicators did more damage than any mathematical weakness. |
| Keeping a machine secret | Machines | The system was published; the settings were the secret. Mostly. |
Section 4 — The Pad
Section page →| Entry | Section | What it promised, and what broke it |
|---|---|---|
| Unbreakable, and almost unusable | The Pad | Key as long as the message, truly random, used exactly once: provably secure and almost impossible to run at scale. |
| What random has to mean | The Pad | Not merely unpredictable to you — unpredictable to everyone, forever. |
| Reuse is fatal | The Pad | Two messages on one pad page collapse the whole guarantee. |
| Distribution | The Pad | The key has to get there first, physically, and that is the whole problem. |
| Destroying the page | The Pad | A discipline that only works if it is absolute. |
Section 5 — Broken
Section page →| Entry | Section | What it promised, and what broke it |
|---|---|---|
| Procedure, not mathematics | Broken | Almost every historical break came from how a system was used rather than from the system itself. |
| The crib | Broken | A guessed fragment of plaintext, and how much it gives away. |
| Traffic analysis | Broken | Who is talking to whom, how often, and how much that reveals without reading a word. |
| Read for years | Broken | Systems broken quietly and left in service, because the break was worth more than the warning. |