Leen Kosten

Leen Kosten was a man with a thorough knowledge of mathematics he often amazed his staff with his knowledge of very diverse mathematical topics. His work at the PTT Laboratory consisted of mathematically analyzing the traffic that passed through the telephone systems of the PTT. When that became too complex he became head of the quite new Mathematical Department. His work on the traffic analyzers was a large part of his doctoral thesis of 1942. He was a very inventive man, constantly full of ideas, to the point that his colleagues occasionally had problems in keeping him from going too far. In a long project like the development of a computer it's impossible to make fundamental changes on a daily basis, so that restrained Kosten. Once a project was solidly founded it lost his attention and he went on to a new interesting subject.

The switching rate of the relays was quite high, and after a while they did not appear to be making reliable contact anymore. The currents through these contacts were extremely small, but Kosten knew the remedy. A small potential was put across the contact by a simple few volts battery. This caused a so-called fritter current across the contact, biting through any eventual buildup of a film of dirt. The DC component could not hurt the signal, as this was a completely balanced AC waveform.

Analyzing Traffic

As traffic analyzer Kosten calculated a lot on the telephone network, but as it grew bigger and bigger, that became more and more complex to the point that other solutions had to be found. The research for a new solution started in 1937. This resulted in the development of special purpose machines that he and his group built from 1940 to 1950. These machines were to imitate the behavior of the telephone network but were never entirely satisfactory as they couldn't be adapted.

Part of these machines were devices that had to simulate the randomness of incoming telephone calls. Solutions were found like using a selector (a turning device to select a free line) with a geigerteller as a simulator. Slowly the news of the ENIAC and new technologies developed for computers that could be helpful for his analyzers changed the traffic analyzers. From 1947 to 1950 he used the computer technologies to add memory to the analyzers to keep track of busy and free lines.

PTERA with Kosten and Van der Poel

Leen Kosten (standing) and Van der Poel working on the PTERA

Designing Computers

When he discovered that the work he did with complicated machines for random input could be simulated with algorithms, his group started working on a real general purpose computer. He also recruited Van der Poel to help with building these machines. His own experience from building the traffic analyzers came in very handy to solve practical difficulties in building the computers.

Together they designed the experimental machine the ZERO to move on to build a real machine, the PTERA. Then for the design of their first commercial computer they went back to what was just a solution at the ZERO, functional bits, and used that principle to design the ZEBRA[1][2].


References

  1. Kranakis, E. , "Early Computing in the Netherlands", CWI Quarterly, vol. 1, issue 4, pp. 61-84, 12/1988.
  2. Poel, W. van der , "The Early History of Computing at PTT", Delft Progress Report, vol. 5, pp. 174 - 184, 1980  .
Pictures: 
Kosten and Van der Poel at PTERAKosten and Van der Poel at PTERALeen Kosten (standing) and Van der Poel working on the PTERA