Communication in the Presence of Adversaries - Burdens of Proof - page 29

When worries over the efficacy of the Enigma led the Germans to switch
to a new four-rotor system in February 1942, Allied cryptanalysts were able
to break the enhanced Enigma by taking advantage of simple errors by
cipher operators: “Once, in December a U-boat cipher operator had care-
lessly let [the fourth wheel] move out of this [neutral] position while
enciphering a message; Hut 8 had noted the ensuing gibberish, and also
spotted the re-transmission of the message on the correct setting. This
elementary blunder of repetition, so easy to make while the Germans held
complete trust in their machines, had allowed the British analysts to
deduce the wiring of the wheel.” 34
Though these cryptanalytic successes meant the Allies enjoyed exten-
sive foreknowledge of German military operations, they were faced with
the serious problem of not showing they knew too much—indeed, how
could the information be used without giving away the breaking of Enigma?
They resorted to performing forms of reverse funkspiel, elaborate disinfor-
mation strategies to provide German military commanders with alternative
explanations for their losses.
More important, they could count on the German's unfailing faith in
the technical merits of their cryptosystems and their corresponding will-
ingness to blame spy networks for the leaks. 35 As Hodges notes, “Bletchley's
continued successes depended upon the willingness of German authorities
to believe that ciphers were proved secure, instead of asking them whether
they actually were. It was a military Gödel theorem, in which systematic
inertia rendered German leadership incapable of looking at their system
from the outside.” 36 Indeed, it was only after the British government con-
firmed the break in 1974 that former German officers came to accept that
their communications infrastructure had been compromised. 37
Perfect Secrecy
In 1917, an American engineer working for AT&T, Gilbert Vernam, stum-
bled upon the perfect cipher while attempting to devise a system to
improve the secrecy of the teleprinter, the printing telegraph. 38 The prin-
ciple was remarkably simple: to represent characters, the teleprinter used
the Baudot code, which assigned to twenty-six alphabetical character and
six nonprinting control characters a five-bit binary code made of either
marks or spaces (2 5 = 32). Thus, the letter a is mark , mark , space , space , space ;
the letter b is space , space , mark , mark , space , and so on. A teleprinter