Just wondering how history questions will fare here. They could be a good addition to our scope.

Pretty much exactly as the title says. Some of the earliest computers were based on valve systems or mechanical arrangements. What was he earliest computer that was not based on either of these systems?

I'm assuming this means a digital electronic computer, but I may be wrong if there's an earlier type!

  • Core rope memory and magnetic core memory were electro-mechanical. Do these count as "mechanical"?
    – Sklivvz
    Commented Apr 21, 2016 at 8:42
  • @Sklivvz I'm going to say yes, but feel free to mention them along with anything else you find.
    – Åna
    Commented Apr 21, 2016 at 8:47
  • 1
    Just to ruin all the fun: en.wikipedia.org/wiki/History_of_computing_hardware Commented Apr 21, 2016 at 8:49
  • @MichaelHampton spoilsport ;)
    – Åna
    Commented Apr 21, 2016 at 8:52
  • @MichaelHampton it's a great resource! Surprisingly, it doesn't give a clear answer to this question, though.
    – Sklivvz
    Commented Apr 21, 2016 at 8:53

2 Answers 2


Since vacuum tube computers are considered fully electronic, and electromechanical computers are evolutionarily a step backward from that, I'll remove them from consideration.

This requires a computer based on the transistor, which was invented in 1947, and still forms the core of every modern CPU, though today they number in the billions.

The first computer to use transistors at all was an experimental Transistor Computer built at the University of Manchester in 1953, with a second version built in 1955. But this computer still had a few tubes left in it. This line of computers never got a proper name.

The first fully transistorized computer, then, was the Harwell CADET, built by the UK's Atomic Energy Research Establishment; it first ran in 1955 but was only in use for four years before being put out to pasture. It ran numerical analysis jobs for nuclear physicists.

At about the same time in 1955 IBM announced the 608 computer, but its claim to being the first is dubious as it does not appear to have actually been on the market until 1957.

  • 1
    Magnetic cores came between vacuum tubes and transistors. In addition to holding bits they could also perform switching tasks, though I don't know how much they were used for actual "computing" purposes.
    – supercat
    Commented Sep 20, 2016 at 20:50

If electromechanical computers count, then it would be remiss to ignore Konrad Zuse's relay-based computers, built just before and during WW2. These were, as I understand it, programmable digital computers, and can therefore be considered alongside the Heath Robinson and the Colossus as the first practical examples of such a thing. I don't believe they were ever used by the German war machine, for which we can probably be grateful.

Though not what we'd now call a programmable computer, the Bombe machines used to help crack Enigma were sophisticated electromechanical calculators, devoted to a specific (and quite vital) task which modern computers can perform only slightly faster. Rather than relays, the active components were rotors with circular arrays of electrical contacts, very similar to the Enigma machine itself but on a much bigger scale. The "menus" they were configured with could be considered programming of a sort.

The oldest still-extant digital computer is the Harwell Dekatron, which uses then-standard Post Office relays for sequencing control, but which however uses cold-cathode and thermionic valves for the high-frequency (about 100Hz) pulse circuit, memory and some calculation circuits. Although much slower than many of its contemporaries due to its heavy reliance on relays (Post Office relays are very big and slow), it was much more reliable and therefore useful, and probably cost much less to build and run.

You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .