Cryptography Reference
In-Depth Information
The method has the remarkable property that the statistical relationship between
successive letter pairs doesn't provide any more benefit during the analysis. (I
don't think that in German, for example, the 1st, 14th, 27th, and 40th letters
of a sentence will still be in a meaningful relationship to each other.) This
distinguishes the method, for example, from the
playfair method
mentioned
in [BauerMM, 4.2.1], which uses only one single square and looks at pairs of
successive
letters. Also, the method avoids pairs of equal letters, since they
won't be converted. To this end, the code writer inserts an 'x':
Atxtila, Sexer.
Of course, that's dangerous. In our case, equal-letter pairs play no role at all
since they are generally not translated into such pairs.
The method described above was broken by the British, by the way, together
with the Enigma code at the famous Bletchley Park (we will get back to this in
Section 2.5.2). A reader of the first German edition brought an article by Noel
Currer-Briggs in [Hinstrip, Chapter 23] to my attention. This article described
cryptanalysis in detail. I'd like to briefly mention a few interesting details.
Just like all methods mentioned so far, this method isn't invulnerable to plain-
text attacks either. The example above (“keine besonderen Vorkommnisse” (no
unusual occurrences)) is rather typical: many messages of this type were sent
by the German Wehrmacht. And when a commander passed an encrypted com-
mand he wasn't likely to omit a 'HEILHITLER' closing ahead of his name.
Of course, the enemy also knew the commander's name (i.e., the last word of
the message). After all, numbers had been stiffly translated into words: '1324 =
ONETWOTHREEFOUR'. That's what we call a
ciphering error
. More about
it in Section 2.5.2.
However, the Germans made it even easier than that for the British. First,
they replaced a blank by an 'X' (another ciphering error). Second, umlauts
were represented as usual, namely 'AE', 'OE', and 'UE'. Since 'E' is the most
frequent letter in German texts anyway, a particularly frequent occurrence of the
'XE' and 'EX' pairs had to be expected. Third, 'J' was replaced by 'II' (more
about the impact of this below). And fourth, there were plenty of long words
like UNTERGRUPPENFUEHRER and GEFANGENGENOMMEN (DEPUTY
GROUP LEADER and CAPTURED) that didn't fit in one cipher line. All these
things helped the cryptanalysts to search for specific digrams in a targeted way,
such as, for example, 'UU' from the word Untergruppenfuhrer split in a 13th
