After Germany (Zuse Z3 and Z4), Sweden was the second country on the European continent to develop its own computer (the Bark), in 1950. The third country to do so was Ukraine (with the Mesm).
The Bark Relay Computer
Little is known about the Swedish Bark (binär automatisk relä-kalkylator, binary automatic relay computer). A few facts: decision to build: December 1948, beginning of construction in February 1949, completion in February 1950, followed by trial runs. Dedicated on April 28, 1950. Commercial use from July 1, 1950. The price for contracted jobs: 3 (British) Pounds per hour of machine time. Operation: 300 days/year, 20 hours per day. The machine originally had 5,000, and later 7,500, relays. The cost of building (with planning, design, construction, and trial runs) was less than US$100,000.
The large-scale binary plugboard-controlled relay computer with floating point notation and conditional jumps was built by the Swedish Telegraph and Telephone Administration under the supervision of the Swedish Board for Computing Machinery (Mathematikmaskinnämndens arbetsgrupp, MNA, Stockholm). Responsible for the project was Conny Palm of the Royal Institute of Technology (KTH) in Stockholm. The principal designers of the parallel machine were the engineer Gösta Neovius and Harry Freese. The Swedish Navy supposedly utilized the Bark for the calculation of artillery trajectories, and the Intelligence Agency for cryptanalysis. In October 1955, the Bark was dismantled for reasons of space.
Who Operated the Bark?
The two contemporary pioneers Goldstine and Willers reported that the Bark relay computer stood in the Royal Institute of Technology in Stockholm.
"The first accomplishment of the group was a relay machine.1 [Conny] Palm and Gösta Neovius built the Bark, binär automatic relä-kalkylator, at the Royal Institute of Technology in Stockholm with funding by the Swedish Telegraph Administration. They were supported by [Carl Erik] Fröberg, G. [Göran] Kjellberg, and a number of other talented people. The design of the machine was due to Neovius and Harry Freese. It was completed in February 1950"2 (see Herman Heine Goldstine: The computer from Pascal to von Neumann, Princeton University Press, Princeton, New Jersey, 2nd edition 1993, pages 349–350 (reprint, first edition 1972)).
1 Much of the material here is contained in a few mimeographed bulletins issued by the Board in its early days.
2 G. Kjellberg and G. Neovius, "The Bark, a Swedish general purpose relay computer," MTAC, volume 5 (1951), pages 29–34.
Willers confirms this site (see Friedrich Willers: Mathematische Maschinen und Instrumente, Akademie-Verlag, Berlin 1951, page 71).
As information from the Royal Institute of Technology (KTH) in Stockholm (personal communications of Arne Kaijser of June 12, 2015 and Johan Martin-Löf of January 4, 2017) reveals, both the Bark and the Besk were installed at the earlier location of the KTH at 95A Drottninggatan in Stockholm (in 1950, ETH Zurich, with the Z4, was therefore the first university in Central Europe with a functional automatic computer).
At the international conference on computing machines in Paris in 1951, several reports dealt with the Swedish computer:
"Renseignements pratiques. – Envisagé et projeté par M. C. Palm qui a ensuite dirigé la construction, le Bark a été projeté en détail par un petit groupe de discussion, dessiné pour la plus grande partie par MM. Freese et Neovius, et bâti aux usines de l'administration royale des Télégraphes et téléphones de Suède. Les premiers plans furent discutés en décembre 1948, et en février 1950 la machine a accompli son premier travail: une table de cosinus. Les frais ont été environ de 400.000 couronnes suédoises, c'est-à-dire environ 25.000 millions de francs français" (see Göran Kjellberg: Quelques problèmes traités avec le Bark, in: Joseph Pérès (editor): Les machines à calculer et la pensée humaine, Paris, January 8–13 1951, Editions du Centre national de la recherche scientifique (CNRS), Paris 1953, page 324).
In English: Practical information – Conny Palm proposed and designed the Bark and then supervised its construction. A small working group planned the computer in detail. The principal engineers were Harry Freese and Gösta Neovius. The machine was built in the workshops of the Swedish Telegraph and Telephone Administration. The first plans were discussed in December 1948, and in February 1950 the machine performed its first task, a table of cosines. The project cost about 400,000 Swedish crowns, i.e. about 25 million French francs (error in the conference proceedings).
The footnote indicates that the costs following the expansion of the memory in January 1951 increased to around 600,000 crowns. Göran Kjellberg was an assistant at the Matematikmaskinnämnden in Stockholm. In his talk, he described the experiences made with the completed machine in February 1950. The trial runs lasted until the end of June 1950. The computer was comprised of 5,200 relays (from January 1951, 8,000 relays).
"Cette machine a été construite par un groupe sous la direction du dr C. Palm.* Elle est constituée autant que possible de matériel standard de central téléphonique automatique, fabriqué par la direction des Télégraphes et téléphones. A présent la machine Bark contient environ 7500 relais électro-magnétiques" (see Stig Ekelöf: Les machines mathématiques en Suède, in: Joseph Pérès (editor): Les machines à calculer et la pensée humaine, Paris, January 8-13 1951, Editions du Centre national de la recherche scientifique (CNRS), Paris 1953, page 136).
* Professeur à l'Ecole polytechnique supérieure de Stockholm (Suède), décédé subitement en décembre 1951.
In English: A group under the leadership of Conny Palm* built this machine [the Bark relay computer]. As far as possible, it was constructed from customary components for automatic telephone switchboards and was built by the Telegraph and Telephone Administration. The Bark currently has about 7500 relays.
* Professor at the Royal Institute of Technology in Stockholm (Sweden). He died unexpectedly in December 1951.
The Besk Electronic Computer
The successor to the Bark relay machine was the Besk (binär elektronisk sekvens kalkylator, binary electronic sequential computer) electronic computer. It was built by the Swedish Board for Computing Machinery in Stockholm under the supervision of chief engineer Erik Stemme. The chief mathematician was Göran Kjellberg. The magnetic drum computer operated regularly from March 1954. Its cathode ray memory was replaced later by a ferrite core memory. The arithmetic unit consisted of flipflop registers.
"The Besk is of very neat construction and is housed in a good-sized room on the second story of the old building of the Royal Institute of Technology (Kunglig Tekniska Högskola) on one of the larger streets of Stockholm. The Bark formerly occupied a room on the first floor of the same building, but nothing remained of it but its framework at the time of my visit in October" (see Nelson M. Blachman: Some automatic digital computers in Western Europe, in: Transactions of the IRE, Professional Group on Electronic Computers, volume EC-5, 1956, no. 3, page 160).
As follows from the travel report of the American physicist Nelson M. Blachman, the Besk was installed on the second floor of the old Royal Institute of Technology building in Stockholm. Its predecessor, the Bark, was on the first floor of the same building, but was already dismantled in October 1955.
Per Lundin sees this differently:
"Besk was installed at MMN, which was located in an old nineteenth-century building at Drottninggatan, the main street of Stockholm at that time" (see Per Lundin: Computers in Swedish society, Springer-Verlag, London 2012, page 87).
Digital computer newsletter: volume 2, 1950, no. 1, page 4 and no. 2, page 4
Digital computer newsletter: volume 8, 1956, no. 1, page 15,Herbert Bruderer: Milestones in Analog and Digital Computing, Springer Nature Switzerland AG, Cham, 3rd edition 2020, 2 volumes, 2113 pages, 715 illustrations, 151 tables, translated from the German by John McMinn, https://www.springer.com/de/book/9783030409739
Göran Kjellberg: Computing machine projects in Sweden, in: Michael Roy Williams; Martin Campbell-Kelly (editors): The early British computer conferences, MIT press, Cambridge, Massachusetts, London/Tomash publishers, Los Angeles, San Francisco 1989, pages 134–137 (conference in Cambridge, 1949)
Gösta Neovius: Activity in Sweden in digital computer field, in: Michael Roy Williams; Martin Campbell-Kelly (editors): The early British computer conferences, MIT press, Cambridge, Massachusetts, London/Tomash publishers, Los Angeles, San Francisco 1989, page 193 (conference in Manchester, 1951)
Göran Kjellberg; Gösta Neovius: The Bark, a Swedish general purpose relay computer, in: Mathematical tables and other aids to computation, volume 5, January 1951, no. 33, pages 28–34.
Herbert Bruderer ([email protected]; [email protected]) is a retired lecturer in the Department of Computer Science at ETH Zurich and a historian of technology.
In my view LEO is an important early computer as explained in the below blog
LEO: The Worlds first business computer
An Important Milestone in the history of computing is the advent of using computers for business and administrative application. That fell to J.Lyons & Co. who were the largest catering company in the United Kingdom in the mid 20th century.
John Simmons, a senior executive of J.Lyons & Co, was responsible for the accounting and office side of the business. With his assistant, T Raymond Thompson they reorganised the accounting and management offices establishing in the early 1930s a Systems Research Office. The business had become so complex by 1947 that they realised they needed some form of automation. A visit to the United States proved fruitless as although there were computers available, none of them had been designed for business purposes. They were given a lead, being told that Cambridge Universitys mathematics laboratory was building an electronic digital stored-program computer, called EDSAC. The technology for this might form a suitable basis for the design of a machine for automating business data.
Simmons and his team were able to gain approval from the Lyons board to provide financial support to Cambridge to complete the EDSAC project. In return, Lyons were to receive technical assistance towards design of their own machine. In 1949 EDSAC was operational resulting in the Lyons board giving the go-ahead for the design and build of their own electronic calculator, for which the name LEO (Lyons Electronic Office) was coined.
A small team of designers and developers led by Dr John Pinkerton were recruited from Lyons and outside companies to build LEO. David Caminer, Manager of the Systems Research Office, was put in charge of the design and programming side.
The first successful live run of a regular business application was in November 1951, the Guinness World Records have acknowledged LEO l as the worlds first electronic business computer. On Christmas Eve 1953 LEO broke another record by running what we believe was the Worlds first computerised payroll. The payslips were for Lyons employees. The success of the Lyons payroll led Fords UK and Kodak to sign up for their payroll to be computerised. Also during the 1950s the Lyons Teashops Distribution project was implemented along with projects for many blue-chip companies and government departments, including the Meteorological Office. The bureau also ran a production control system for a clock manufacturer. All this using LEO which had a 2048 17-bit word memory.
The success of LEO l led to Lyons launching LEO Computers Ltd. as a wholly owned subsidiary. They went on to develop LEO ll and LEO lll computers. In addition to their use by Lyons, nearly 70 organisations purchased these LEO machines, including British Oxygen, Shell Mex & BP, Dunlop Rubber. HM Dockyards purchased 4 LEO IIIs and the General Post Office purchased 14, the last of these being decommissioned in 1981.
Photo of LEO l cabe found at: https://www.leo-computers.org.uk/wp-content/uploads/2020/09/1-b-4.jpg