For Enigma fans

[Clive D.W. Feather]

David Hansen said:
>> I was under the impression that the bombe was not "based on" the bomba,
>> but was an near-exact duplicate.
> It was very much simpler and relied far more than the bombe on reducing 
> the problem to manageable levels by "hand", before putting the problem 
> to the machine. Imagine something the size of an ice cream box mounted 
> on a tricycle or hand cart. ISTR bomba was a brand of ice cream.

Yes: the machine was named after it because the idea came to the inventors
while eating one.

The bomba consisted of 6 interconnected Enigmas. It solved a specific
problem: in the early days, an operator would pick two three-letter
indicators, say ABC and DEF. He would set part of the machine to ABC, then
encrypt DEFDEF, getting (say) PQRJKR, and transmit ABC PQR JKR as the
message header. He would then re-set the machine to DEF before encrypting
the message.

The bomba solved the specific case where you had three messages with
indicators of the form:

    --- X-- X--
    --- -X- -X-
    --- --X --X

*and* X was not steckered. It became useless when the number of steckers
went up from 4 to 10, and with the introduction of the fourth and fifth
wheels.

> The bombe didn't spit out German but came up with possible settings to 
> try by "hand".

The bombe tested a crib: a guess that a certain sequence of letters in a
message corresponded to a certain piece of plaintext. It would report if
there was a position of the Enigmas consistent with that guess and, if so,
one possible stecker link. Someone could then take that information to an
Enigma (or modified TypeX) and see whether they could reconstruct the rest
of the steckering (or find that it didn't work).

Each of the three blocks on the bombe was a separate unit; since there were
5 wheels to choose from and 3 slots to put them in, there were 60 possible
arrangements of the wheels to test. That required 60 units, or 20 different
bombes, all testing the one crib. The three wheels at the right hand end of
the centre row were not connected to anything else - they were a separate
Enigma used by the operator to test a guess.

> Once the correct settings had been worked out the bombes 
> were used for the next job

Not quite. Once the stopping position had been written down, the bombe was
re-started in case it found another position that also worked. When it had
gone through the complete 26x26x26 cases, another job was set up on it.

> while the German was produced on British 
> TypeX cipher machines modified to emulate an Enigma.

Right.

> IIRC changes introduced by the Germans in the late 1930s made the bomba 
> fairly useless, unless very lucky.

See above.

-- 
Clive D.W. Feather          | If you lie to the compiler,
Email: clive@davros.org     | it will get its revenge.
Web: http://www.davros.org  |   - Henry Spencer
Mobile: +44 7973 377646