I have compiled a list of all the DEUCE documentation I can track down, with links to copies on the Web. If you can provide any additions to the list, please contact me, David Green, at . I have also posted scans of several of the more interesting documents held in The John Rylands University Library at the University of Manchester, with the kind assistance of the archivist Dr Barry White, and of many documents made available by David Leigh. They can all be accessed from the list. The English Electric Company developed several second generation computers to supercede DEUCE, including the KDN2, KDP10 and KDF9. The KDF9 Programming Manual and User Code Digest are on the Web.
The origin of DEUCE traces back to 1945 when Alan Turing and a small team of specialists started designing a computer at the National Physical Laboratory in Teddington, England. Some of the correspondence from around that time can be found in The Turing Archive for the History of Computing at AlanTuring.net.
In time, this computer came to be known as the Automatic Computing Engine (ACE).
To test their ideas, the team first built a small version called the Pilot ACE. It successfully ran a simple program in May 1950, and went into full time service in 1952. Much of the Pilot ACE has been preserved and is now displayed in the London Science Museum.
The English Electric Company (a large engineering Company in those days) had contributed to the research by providing a team of about a dozen engineers and technicians to help at NPL. They decided to build a commercial version of ACE. They called it DEUCE.
DEUCE stands for "Digital Electronic Universal Computing Engine". I suspect they called it DEUCE because it followed ACE, then thought of words for the acronym afterwards.
The first DEUCE was delivered in the spring of 1955. Like the Pilot ACE, DEUCE used mercury delay lines backed by a magnetic drum for storage and used punched cards for input and output, but the physical construction was improved to give better reliability and easy maintenance. Thereafter, EE upgraded DEUCE in several ways. Within a few months they introduced the Automatic Instruction Modifier (AIM). Later they made it possible to read 64 columns per card rather than 32 (and hence two instructions per row rather than the orignal one instruction per row). Later still, they increased the memory size by the addition of seven extra delay lines (which made programming even more complicated, see here).
For more technical details, particularly about UTECOM (the DEUCE installed in Sydney, Australia), see Robin Vowels' web site.
According to Lavington, the company sold about 31 DEUCE Mark I and DEUCE Mark II machines between 1955 and 1964.
I have a soft spot for DEUCE. It was the computer on which I learned to program. However, I think DEUCE has a fair claim to being, for programmers, the most complicated computer ever put into general production. See for yourself. The programming manual is available at John Barrett's Web site. Or check out an example of some program code - the solution to a simple programming exercise given to new chums at Kidsgrove: specification (44kB), flowchart (261kB), coding sheet (551kB).
Other documents describing how to use DEUCE include: "Principles of Programming", "Programmes to Aid Programmers", "DEUCE Control Panel Manual", and "Standard Operating Instructions for DEUCE".
John Barrett has a very interesting web site dedicated to DEUCE, with emphasis on the people who worked with the machines.
Last updated Wednsday, 12 July 2006.