Slashdot Mirror


Was Zuse's Z3 the First Programmable Computer?

Roland Piquepaille writes "Several years before the Colossus in the U.K. and the ENIAC in the U.S., the Z3, built by Konrad Zuse in 1941, was crunching numbers in Germany. In a short article, the Register reports on allegations that the Z3 was the first programmable computer. Based on a binary floating-point number and switching system, it had all the attributes of today's computers, such as a control block, a memory, and a calculator. But it didn't have the ability to store the program in the memory together with the data because the memory was too small. It had a 64-word memory of 22 bits each and was able to handle four additions per second and to do a multiplication in about five seconds. And it was pretty big: five meters long, two meters high, and 80 centimeters wide. It was destroyed during WWII, and later rebuilt in 1960/1961. You'll find more details, pictures and references in this analysis of this ancestor of modern computing. [Additional note: you can find other references to the Z3, Colossus and Eniac computers in this former Slashdot item, posted in October 2000.]"

6 of 450 comments (clear)

  1. Also claimed by... by LV-427 · · Score: 4, Informative

    The ABC Computer at Iowa State University, by John Antasoff and Clifford Berry.

    1. Re:Also claimed by... by Tar-Palantir · · Score: 5, Informative

      The ABC was not really programmable (it lacked control structures), it was more of an automatic calculator than a computer. It was also slow, error-prone, and had a ridiculous output system involving burning (!) holes in paper cards.

      A nice book talking about the early development of computing in the US (so no Z3 or Colossus, sorry) is ENIAC, by Scott McCartney. As the title implies, it's largely about the ENIAC, but ABC is given some treatment as well (particularly in contrast with the far more advanced ENIAC).

  2. Re:This is not a computer.... by Wyzard · · Score: 4, Informative

    According to the article, the program was stored -- on punched film. It couldn't store the program in RAM so it would just read instructions from the film as it came time to execute them, but that doesn't make it any less a stored program.

  3. Zuse's first design surfaced in 1936... by Aphrika · · Score: 4, Informative

    Or at least the plans for the Z1 did. IIRC he tried to get it built, but the engineers thought he was a conman. He eventually got it completed in 1938.

    The next model, the Z2 was partly finished before Zuse got conscripted into the army, obviously they were oblivious as to the importance of his developments.

    Incidentally, it's important to point out that although the Z3 had government money behind it, it was built and used by Zuse personally at home to solve problems with wing flutter for Heinkel where he worked. It was destroyed by chance when his home was hit in a bombing raid.

    Zuse also developed the first multi-purpose computing language 'Plankalkul' too. Quite an impressive achievement for a mathematician who developed a computer simply to enable him to do his wing calculations more effectively.

  4. Re:This is not a computer.... by mark-t · · Score: 4, Informative
    You are confusing "Computer" with "Von Neumann Architecture".

    Whether or not there is a stored program does not affect whether or not it is a programmable computer.

  5. An overview of contenders to the crown. by arevos · · Score: 4, Informative

    First let's start with ENIAC. ENIAC used valves, was electronic, was Turing Complete, and was designed to be Turing Complete. Which means that it could, theoretically, solve any problem currently solvable by today's machines (given enough time). Because it was Turing complete, it was obviously programmable.

    The Z3 used mechanical relays instead. If I recall right, the Z3 could be Turing Complete with a little hack. In 1998, if I remember right, someone showed that conditional jumps could be implemented by quite literally forking the punched tape that was fed into it. So the Z3 was Turing complete, but wasn't quite designed to be. It was, however, quite programmable.

    Collosus wasn't Turning Complete, but it was damn fast for what it did. It was programmable, and used valves like ENIAC later did.

    Thus, the Z3 was the first Turing Complete (sort of) programmable computer ever made.

    Collosus was the first fully electronic, programmable computer. It was also the first programmable computer used to break encryption.

    ENIAC was the first computer designed to be Turing Complete.

    Strongest contender to the title of the first "real" computer is, in my opinion, the Z3.