E-mail from and to <a href="mailto:Frode.Weierud@cern.ch">Frode Wierud</a> concerning the Enigma Machine. 
<p>
<p>
Hi Edward and Jon,
<p>
I have just found your nice Web page "Learning about The Enigma Machine 
and other examples of Cryptography and Mathematics". I have now made a
link to your page from my own Cryptology page at:
<a href="http://home.cern.ch/~frode/crypto/index.html">
http://home.cern.ch/~frode/crypto/index.html</a>
<p>
I have not checked the contents of all your pages for accuracy but I have
found one mistake that you ought to change. It appears on your page
"Enigma Machine Description" and has to do with the period of the Enigma
machine. You say the period is 17576 while in reality it is 26x25x26 =
16900. This is due to a mechanical anomaly of the Enigma stepping
mechanism. When the third (leftmost) wheel steps it will also move the
middle wheel at the same time. The reason has to do with the V shaped form
of the notch which means the stepping pawl will push as well on the edge
of this notch and move the middle wheel together with the third wheel. A
detailed explanation of all this can be found in David Hamer's Cryptologia
article which you can find at his Web page at:
http://www.eclipse.net/~dhamer/Downloads.htm (dead link)
<p>
The title is: "Enigma: Actions Involved in the 'Double Stepping' of the
Middle Rotor"
<p>
Best wishes,
        Frode
<p>
<p>
Frode Weierud                
<p>CERN, SL,  CH-1211 Geneva 23, Switzerland                     
<p>Phone  : +41 22 7674794
<p>WWW    : wwwcn.cern.ch/~frode -- dead link
<p>Fax    : +41 22 7679185
<p>E-mail : Frode.Weierud@cern.ch
<p>
<p>
<br>
On Thu, 25 Mar 1999, Edward Aboufadel wrote:
<p>
      
      I've heard about the period issue for the Enigma machine that you 
      discussed.  My reading of various sources last year was contradictory, 
      as some reported what you said and others reported the 26^3 period.  I 
      am now thinking the right thing to do is to include both ideas, with 
      references.  Maybe if I ever get a close look at a real machine ...
      
<p>
Dear Edward,
<p>
You can rest assured that the 16900 period length is the only correct
answer. I have had access to many real Enigma machine and the stepping is
as I described. It is only the Abwehr model of the Enigma machine which
have a different stepping. The stepping of the normal Enigma machines
are usually called Enigma stepping, while the one used on the Abwehr model
goes under the name cyclometric stepping.
<p>
Where the confusion comes in is with regards to the number of starting
positions. It is clear that the machine can be started in all of the
26x26x26 = 17576 positions but that from each position you will only get a
full cycle of 16900. David Hamer have explained about the run-in tails in
his article.
<p>
The Cipher Simulation Group has done extensive simulations of all of
Enigma machines as well as other cipher machines. You can also download a
few public domain versions of our Enigma simulators from the CSG Web page
at: <a href="http://home.cern.ch/~frode/crypto/CSG/index.html">
http://home.cern.ch/~frode/crypto/CSG/index.html</a>
<p>
Best wishes,
        Frode