The slide show combines various accounts in books generally available with new information released more recently. It attempts to portray Turing as a gifted man who found himself in an environment, at Bletchley Park in particular, where his particular skills and abilities, as well as his understanding of what was to be computer programming, were highly valued. The contention is that at Bletchley Park and in certain computer development work afterwards, Turing was able to perform as a specialised worker and at his best

1. The World of Alan Turing, or
From the Turing machine to the first
commercially-available
(general purpose) computer

2. The paper “On computable
numbers…” (1936-7)
• This paper introduced 3 (principal) ideas or
concepts – the Turing machine, with its read-
write head, 4 basic commands and its tape as the
‘memory’
• The machine description, which Turing refers to
as the m-configuration (its collection of – finite –
states, thinking in terms of automata)
• The idea of the machine which can read – and
operate on – its own description… the universal
Turing machine (as a way of tackling the
Entscheidungsproblem)

3. Look first at the Machine and its m-
configuration - from Marvin Minsky’s text
“Computation”
Turing expresses the m-
configuration as a table in
which the rows are ‘action
descriptions’, e.g. (A)print
0; move 1 square to the
right
(B) Move 1 square to the right
(C) Print 1; move 1 square to
the right
(D) Move 1 square to the right
and go to (A)
Printing 010101…..

4. Now look at the notion of the
Universal (Turing) machine
• This is a machine which can read a series of m-
configurations and thus ultimately operate to any
specifiable computational sequence (an
algorithm)
[ of course, not all such algorithmic operations are
deterministic – i.e. produce a defined
result, which is repeatable ]
The importance of this notion is its universality – it
is a way of ‘interpreting’ any expression of an
algorithm (and thus in our terms, any program)

5. How did Turing become involved in
cracking the Enigma code?
• He initially built on the work of the Polish team
led by Marian Rejewski, who had catalogued
patterns based on the first 6 letters of
intercepted messages:-
Enigma messages start with the 3 – letter message
key & its repetition, for example the sequence L
O K R G M, in which L and R encrypt the same
letter, the first in the message-key. This tells us
that L and R are related by the initial setting of
the Enigma machine

6. If you have 4
messages, starting with • If enough messages are
the 6 letters of repeated received in a day, patterns
key -
Message 1 - L O K R G M
can be discerned, for
Message 2 - M V T X Z E example if you have a
Message 3 - J K T M P E complete set of 1st and 4th
Message 4 - D V Y P Z X letters –
ABCDEFGHIJKLMNOPQRSTU
Now we can see how V W X Y Z and
letters are encoded in
these messages –
FQHPLWOGBMVRXUYCZITN
JEASDK
1st letter
Now we see that A is linked to F; look
ABCDEFGHIJKLMNOP…
up F on the top row, it is linked to W;
look up W on the top row, it is linked
4th letter
to A – which is where we started.
P M RX

7. Think now of how the Enigma machine
is constructed -

8. The Enigma shown has 3 scramblers
(rotors)
• If we think of the relations previously described
between the 1st and 4th letters, we see a ‘chain’, here
- in the simplest case – A->F->W->A (with 3 links)
• It can be shown that the number of links in these (multiple)
chains is wholly dependent on the scrambler settings:
Rejewski’s team catalogued the chain lengths generated by
each of the 105,456 scrambler settings. Now each day
messages were received, they could see the first 6 letters
and thus identify the chains and the scrambler settings that
created them. Thus the scrambler part of the day-key has
been separated from the extra encoding performed by the
plugboard, which operates as a straightforward substitution
(e.g. R may be plugged to L), operating each time a certain
letter is encyphered.

9. When Rejewski’s team met Knox &
Turing
In July 1939 in a hut in the Pyry forest outside
Warsaw, the French intelligence service
arranged a meeting between the British cipher
service and the Polish team. This meeting
resulted in the handing over of a
reconstructed Enigma, a number of bombes of
Polish design, and the basis for cracking the
Enigma messages using a 3-rotor Enigma.

10. Issues
While the Poles had devised a method for cracking
Enigma, it is perhaps important to note that a major
reason for the meeting was an increase in the design
complexity of message coding by the Germans –
• 2 more rotors so that 3 were chosen from 5
• No. of plugboard cables up from 6 to 10
The Polish methods could not handle these changes!

11. Start of operations at Bletchley
Park (“phoney war”)
When Turing joined Bletchley Park, others like
Gordon Welchman had already arrived and
were preparing the ground for the
codebreaking work. There was a heirarchy
being built up to provide support & services to
interact with the codebreakers & lots of
others, some known to Turing, became
involved. Among these were Max Newman,
his mentor at Cambridge in much of his work.

12. Members of the initial heirarchy at
Bletchley Park

13. The different Enigma designs –
Abwehr, Kriegsmarine, Luftwaffe
The German army, navy and airforce used Enigma
designs of varying complexity –
• The army & airforce used 3 rotors chosen from 5,
• The navy 3 or later 4 from a total of 8
The teams at Bletchley were able to develop
techniques in Bombe design & usage to handle
army/airforce and navy messages up to the first battle
of the Atlantic. However, as U-boat usage of codes
and keys became more sophisticated, decoding
difficulties were increased and new strategies and
techniques were needed.

14. The problems come to a peak
In 1942 and early 1943, there was an 8 month
break in being able to solve Enigma messages
from the navy and the U-boats –
• Attacks in the Atlantic were intensifying
• 4-rotor Enigmas had been introduced
The technique known as ‘gardening’ was the
only real answer Bletchley had. A new strategy
was needed. The study of ‘Fish’ was vital.

15. ‘Dilly’ Knox dies, having broken the
Abwehr Enigma
• Knox was a scholar & academic, not a mathematician
(but he was originally attached to Admiralty room
40). He suffered from cancer and died in Feb. 1943.
• He was determined to create a break into the
Abwehr Enigma so that the Double-Cross messages
sent via turned German agents could be verified as
being ‘believed’ in Berlin. He succeeded in Dec. 1941.
• Once it was known that Double-Cross worked,
continuing to read Abwehr Enigma messages became
fundamental.

16. The Lorenz messages (“Tunny”)
The German high command centres used
teleprinter-based messaging, not Enigma.
Messages were encoded and sent using
teleprinter code from 2 centres, on various
routes manned by Axis forces. It was these
messages that determined the movement and
disposition of troops, and following the losses
in the Russian and desert campaigns they
became vital to intelligence. They were
encoded using the Lorenz ciphers.

17. How was Lorenz different?
• There were 3 Lorenz designs, but in
general, Lorenz used twelve wheels or rotors
in two groups. 5 were used separately to
encode, 5 more could be used either in unison
or in tandem to break up the message
• There were 2 motor wheels which controlled
the two groups of 5
• The encoding wheels were not connected at
every letter and the inter-rotor connecting
circuits were varied frequently.

18. The machines used to break Lorenz
ciphers
Initially, there was Heath-Robinson, a machine
which read 2 tapes, one the encoded message
and the second an example decode based on
a chosen setting of the encoding wheels.
Because tapes were read at a maximum speed
to try and solve the complexity of having to try
many wheel settings, this machine had many
problems with breaking and tearing tapes.

19. Enter the Post Office research
station team
At Dollis Hill the GPO had an engineering team
developing electronics for working on advanced
exchange design –
• Valve electronics so reliability issues, but
• Circuitry design was advanced and offered complex
functionality
Tommy Flowers, head of the Dollis Hill team, was
certain a machine could be designed to overcome the
Heath-Robinson problems: Max Newman & Donald
Michie are generally credited with the overall
Colossus design

20. Colossus I
• 1500 valves, reads only 1 tape and stores the
encoded message within machine ‘memory’
• Uses algorithms to test possible wheel-setting
combinations to analyse decodes of the
message statistically
• Results can be printed to enable the most
likely settings on which to concentrate efforts.
Delivered to Bletchley December 1943.

21. Colossus II and the timing of efforts
• 2500 valves bringing more memory & power
• 5 different encodings can be processed
together – stops breaks when no decodes.
Combined the advantages of the original design
-> The most successful design – eventually 10
were delivered!
The new Colossus was first delivered a few days
before D-day and this allowed 2 simultaneous
attacks – disinformation and codebreaking.

22. Colossus design - modular

23. War ends – what happened next? Bletchley Park was closed and there
was a bonfire to avoid secret information being retained.
The computing research efforts continued in 4 or 5 groups, some of
which had already started during the War – Radar at
Malvern, NPL, Manchester, Cambridge, others
TOWARDS THE FIRST STORED-
PROGRAM COMPUTERS – ON SALE!

24. Reviewing the (unfinished) story

25. Main development groups in 1946
• NPL where Turing began initially after the War
– attached to atomic energy research site at
Harwell, very bureaucratic, already secretive
• Manchester University where memory
advances came together and where Newman
started the design lab. – birth of the 3 registers
(A, B, and the combined C- PI & program
counter). Turing joined him later.
• Cambridge under Hartree and then Wilkes – a
pragmatic approach based on human
“computers”- produced EDSAC and then LEO.

26. The different approaches
• NPL was where Turing originated the design
for the ACE – not commercially available till
after he died
• Manchester was where the first stored-
program computer actually ran a program- an
advance impossible without perfecting the
Williams-Kilburn tube used as a program store
• Cambridge started after the Moore school
work and chose a more pragmatic approach

27. The Manchester “Baby” and the
advances that followed there
This basic “fetch cycle”
design was varied and
improved on several
times at Manchester
University computing
lab. – storage was
increased (to 8k words)
to provide more
flexibility and allow
different usage strategies

28. The Moore school ideas & Maurice
Wilke’s work at Cambridge
• Cambridge computing lab. started in 1937 with
a differential analyser & desk calculators.
• After the war Wilkes got J. von Neumann’s
“Draft report on the EDVAC” and later went to
a Moore school course. The Moore school had
built ENIAC but it only worked in Nov. ’45.
• Wilkes’ computer- EDSAC - used the same
delay-line technology as ENIAC- 4 to 32 delay
lines, 3000 valves. Printed a squares table 6
May 1949.

29. Turing after 1948-49 and his
changing interests
• After the practical successes at Manchester in
1948, Turing became a thinker and a
documenter again – he created the
programming manual for the Manchester
Mark 1, later the Ferranti Mark 1
• He moved to thinking about AI and machine
learning and how machines, and abstract
‘minds’ might re-design themselves – resulting
in the “Turing test” paper of 1950

30. Source books & papers
1. Campbell-Kelly, Martin & Aspray, William, Computer: A History of the
information machine, Basic Books, New York, 1996
2. B. Jack Copeland & others, Colossus: The Secrets of Bletchley Park’s
Codebreaking Computers, Oxford University Press, 2006
3. Hinsley, F.H. & Stripp, A., Eds.: Codebreakers: The inside story of Bletchley
Park, OUP, 1993
4. Lavington, Simon H. Early British Computers: The story of vintage
computers & the people who built them, Manchester University Press
1980; and A History of Manchester Computers, British Computer
Society, Swindon, 1998
5. Singh, Simon The Code Book: The Secret History of Codes &
Codebreaking, Fourth Estate, London 1999
6. Leavitt, D. The man who knew too much: Alan Turing & the invention of
the Computer, Phoenix, London, 2007
7. Smith, Michael, Station X: The Codebreakers of Bletchley Park, Pan
Books, London, 2003