One of the books I'm writing will have the heroine and the hero for that matter, working on breaking a difficult French cipher. I'll have them work on something based on Jefferson's disks
(Image from wikipedia)
Given the technology of the day, this would have been very difficult to break, and even today if used correctly would give the NSA fits. (They'd still break it, but it could be made harder if it were used as a superenciphering step.)
That said, I'd like to set the record straight. According to "The Man Who Broke Napoleon's Codes, the story of George Scovell" by Mark Urban, the French weren't that sophisticated. They actually had in their "Le Grand Chiffre" a sophisticated two-part code of 1400 or more groups. It possessed redundancy for common phrases and a number of similar good features. It was certainly better than the "Playfair" ciphers the British would adopt in the 1830's.
The trouble is the French clerks were lazy, and laziness, while a virtue in computer science, it is a disaster in security. They would only encipher parts of the message and the context would give Major Scovell all sorts of clues to what the rest of the message meant. Things like "after the battle at 872 last week general 8.5.6...." Needless to say, they would have been much better off either encrypting the whole thing or sending it just in plain.
Then that's why I'm blogging in English and not Francais.