[syhprum (OP)]

Many successes in decoding enigma coded communications were due to bad operating procedures what was the correct way to use the machines ?

[CliffordK]

Wikipedia has an article about cryptanalysis of the enigma, including operating shortcomings

It sounds like a major shortfall was thinking that the code was uncrackable, and perhaps not having better intelligence to determine how much progress other countries had made on cracking the enigma.  Otherwise they would have taken more care in the coding.  It sounds like one major shortfall was to repeat messages verbatim on both high encryption and low encryption networks.  Also the use predictable message headers and word repetition on some messages. 

One might expect separate keys to be used for point-to-point communications, and point-to-group communication, perhaps defining various subgroups.  So, a message intended only for U-Boat XYZ could only be deciphered by U-Boat XYZ.  However, a message sent to a fleet would use a key specific to that fleet.  Each operator sending messages specifically to Berlin would use their own key.

[syhprum (OP)]

The "Wiki" problem, although  it is always interesting to gather the views of NSF correspondents there seem to be very few scientific subjects that have not been covered by Wiki.
The whole Enigma story seems a case in point and every one of my questions seems to have been covered by the article quoted.

[CliffordK]

Unfortunately I don't have an enigma machine in my basement, and haven't taken apart the mechanical calculator yet.

I wonder how many of these shortfalls still occur today.

Certainly the use of commercially available cryptography software would allow hackers access to the program, and perhaps even the source code of the cryptography system.  Somewhat like the Germans basing their secure encryption on a commercially available enigma.

It is also unclear the changes made from one enigma generation to the next.  It sounds like a lot was based on adding minor mods to the previous generation.  From a technical standpoint it makes sense to build on what one already has. 

When the Germans added wheels or other features, did they reuse the old wheels, so 90% of the machine was identical to the previous version?  Thus the allies only needed to crack the new features, rather than having to crack a completely new device.

[syhprum (OP)]

I see a simile between breaking of enigma code and the development of atomic bombs, the Germans knew that it could be done if enough resources were poured into it but were incredulous as to how much effort the allies were prepared to put into it.
 

[CliffordK]

Had the allies not sequentially cracked enigma codes, then new updates may not have been distinguishable from previous versions, and would have resulted in confusion. 

Apparently the Germans encrypted just about everything, while the allies used the encryption judiciously. 

The Germans used a decentralized decryption groups, while the British decided to put all their eggs in one basket. 

Would WWII have been different had the Germans discovered and bombed Bletchley Park?

[evan_au]

When the Germans added wheels or other features, did they reuse the old wheels, so 90% of the machine was identical to the previous version?

Adding a few new wheels makes much more than a 10% change to the behaviour of the machine.

six plugboard leads, leaving 14 letters unsteckered. January 1939 when the number of leads was increased, leaving only a small number of letters unsteckered.
three different rotors for the three positions in the scrambler. (This continued until December 1938, when it was increased to five and then eight for naval traffic in 1940.)

By delivering 4 extra leads, or 1 extra wheel, the number of permutations increases dramatically
The 3-rotor scrambler could be set in 26 × 26 × 26 = 17,576 ways, and the 4-rotor scrambler in 26 × 17,576 = 456,976 ways.
With six leads on the plugboard, the number of ways that pairs of letters could be interchanged was 100,391,791,500 (100 billion) and with ten leads, it was 150,738,274,937,250 (150,738 billion).

In delivering 2 extra wheels, the internal wiring of these new wheels was initially unknown, and each wheel could be set in 26 initial positions. So this prevented decryption of the traffic on most days, until the circuitry of the new wheels was deduced.

With the huge number of possible encryption codes, even small hints like "no rotor should be in the same slot in the scrambler as it had been for the immediately preceding configuration" significantly reduces the number of combinations to be tried.

Similar scaling applies with modern cyphers - if you make the prime numbers in public-key cryptography 10% longer, the code becomes millions of times harder to crack, not 10% harder.

[CliffordK]

Adding a few new wheels makes much more than a 10% change to the behaviour of the machine.

In delivering 2 extra wheels, the internal wiring of these new wheels was initially unknown, and each wheel could be set in 26 initial positions. So this prevented decryption of the traffic on most days, until the circuitry of the new wheels was deduced.

True,
They may have chosen to preserve backward compatibility. 

The basic wheel & cog system remained unchanged, or was sequentially modified.
Had they made 5 unique wheels rather than reusing old wheels, the problem would have been inordinately more complicated to decipher. 

How does the system work?

Wheel 1: rotates forward one notch once per letter.
Wheel 2: rotates forward one notch on letter 26.
Wheel 3: rotates forward one notch after 26x26 = 676 letters.
Wheel 4: rotates forward one notch after 26³ = 17576 letters. 
etc.

That would mean that the 5 cogs would hardly do more than 3 cogs, but primarily allow more changes for the first 3 cog positions. 

Otherwise, the 4^th and 5^th wheels would be somewhat like hardcoding the patch panel on front.

It would have also meant that given 3 known wheels and 2 new wheels, one could essentially decode one new wheel at a time.

Perhaps they should have revised the wheels by adding 3 umlauts for 29 characters.  [xx(]  Especially beneficial if the typists would randomly either leave umlauts off of words, or add them in where they weren't supposed to be, assuming that could be done without affecting the meaning.