422 | 38 BANBURISmUS REVISITED
There was already some playful language here. In 1928 the Bell Telephone Laboratories invented
a measure that was the base-10 logarithm of the ratio of two measures of anything that could
be measured, and named it the ‘bel’ in honour of their founder Alexander Graham Bell; they
further named the more practicable one-tenth unit the ‘decibel’. So soon after the end of the war,
Good was unable to go on to say that Hut 8 extended their Banbury vocabulary by re-naming
the bel in their context as the ‘Ban’ and their one-tenth and one-twentieth scoring units as the
‘deciban’ and ‘half-deciban’ or hdB. As Christine Ogilvie-Forbes explains, ‘decibanning’ then
entered Hut 8’s vocabulary as the name of a procedure using ‘large sheets of paper and simple
arithmetic with the odd Greek letter thrown in’.^10 She recounts:
I was decibanning the day Churchill came round. He leaned over me and said ‘where was the
message?’ No letters in sight. Alexander rather shortly said it had nothing to do with the mes-
sage, only maths. The men had lined up to be introduced and asked by Churchill what they
had been previously—‘undergraduate, undergraduate etc.’ Alexander said ‘draper’. That slightly
checked the flow.
Alexander had been Research Director of the John Lewis Partnership before being recruited
to Bletchley Park. ‘Draper’ was his amusedly diminishing way of referring to one of the United
Kingdom’s biggest retailers of all kinds of cloth.
Chains and depth cribbing
Enigma’s complexity arose from the choice of three wheels and their order, their three ring
alphabet positions, and the plugboard pattern. As described in Chapter 12, the bombe could,
given time, find what combination of these would decipher a message text so as to match a
crib with which it had been provided. The plugboard pattern changed daily, while the wheels
and ring settings changed every two days. The job of the Banburist cryptanalysts was to speed
the bombes’ work by reducing the number of wheel orders to be tested. No two situations
facing them were ever the same. They started each time with the whole assembly of plausible
or better alignments of pairs of messages resulting from the two processes already described
in Chapter 13. Each alignment, with its credibility label attached, provided the alphabetic
distance between a pair of enciphered letters in the enciphered message setting (Chapter 10).
Starting with the most credible alignments, and linking pairs with one letter in common,
they constructed chains with more enciphered letters spaced out according to the known
distances.
These relationships were only relative. To find the true relationships of the enciphered letters
to their underlying clear (i.e. plaintext) equivalents, each chain of letters spaced out by gaps was
tested in each of the twenty-six possible positions against a true alphabet. Enigma’s reciprocal
feature immediately doubled the number of letters along the chain (for if A was enciphered to
Q, then Q was enciphered to A) and correspondingly diminished the number of gaps. Where
this produced contradictions, that speculative positioning of enciphered letters against true
ones was eliminated.
For those chains and positionings which survived the first test, some further alignments with
next-best credibility were brought in to add evidence. As gaps were filled and the twenty-six
tests were repeated, further contradictions led to further eliminations. Eventually a particular