I would like to know what is considered to be the first electronic digital computer.

A literature ambiguous on this. I found these claims:

  • ENIAC - a computer constructed by Mr. Eckert and Mr. Mauchly at the University of Pennsylvania
  • Colossus - a computer used in Bletchley Park to break the Enigma code, connected with Alan Turing (correction provided in comments: in fact Colossus was used for breaking Lorentz SZ40 enciphered teleprinter)
  • Harvard Mark I - a computer constructed in collaboration between Howard Aiken and IBM
  • Z1, Z2 and Z3 - a computers constructed by German scientist Konrad Zuse

Personally, I would say that ENIAC is the right answer, because Colossus was single purpose electromechanical machine, Harvard Mark I was rather an electromechanical calculator than a computer, and Zuse's computers were highly experimental devices (Z1 and Z2 were electromechanical, Z3 was never completely constructed).

ENIAC was completely electronicised programmable universal computer. But on the other hand it used decimal coding of numbers instead of binary one. If we take this into account, the first computer should be EDVAC or BINAC.

Could anybody please shed more light on this? Some resources would be appreciated as well.

  • 10
    $\begingroup$ Why on Earth would decimal encoding disqualify ENIAC from being a computer? If it's not a computer, then neither are a wide range of IBM devices from the 1950s and early 1960s. $\endgroup$ – Mark Apr 13 '20 at 2:10
  • 5
    $\begingroup$ A similar question was asked on retrocomputing.SE nearly 4 years ago. You might find other relevant information there, too. $\endgroup$ – user7761803 Apr 13 '20 at 8:19
  • 1
    $\begingroup$ Indeed. It's not strictly necessary for a general purpose computer to use binary; this is rather an effect of basing the design of the computer on the transistor. But if using transistors (and binary) is your criterion, then definitely check out the Retrocomputing question linked above. $\endgroup$ – Michael Hampton Apr 14 '20 at 11:06
  • 2
    $\begingroup$ According to Zuse the Z3 was completed, but destroyed, and only reconstructed in the 60s. What for some may disqualify the Z3 is that it was never intended as a turing complete machine, and thus it was only discovered in 1998 that it indeed was, if only through a series of convoluted tricks. But you are seeing this through the eyes of today, you should really see it through the eyes of back then, which would make all these machines electronic computers, as they were usable for a wide variaty of things, and not only a single task they were constructed for. $\endgroup$ – PlasmaHH Apr 14 '20 at 19:10
  • 1
    $\begingroup$ "Colossus was [a] single purpose electromechanical machine ... used in Bletchley Park to break the Enigma code," Some (understandable) confusion here. Enigma was broken (in the latter part of WWII) by Bombes, which were electromechanical. Colossus was valve/tube based and was used to break TUNNY (Lorentz SZ40 enciphered teleprinter). Colossus was digital and electronic but not stored program control nor, ironically, Turing complete. $\endgroup$ – Graham Nye Apr 17 '20 at 22:00

The literature is ambiguous because the question is ambiguous. The characteristics of these early machines are well known, but one can set the criteria for being "electronic computer" according to their preference. Does it have to be completely non-mechanical? Turing universal? digital? binary? History can tell us what happened, not what choice of words we should prefer.

One early candidate missing from the OP list is the Atanasoff-Berry computer, which Wikipedia calls "the first automatic electronic digital computer". It was conceived by Atanasoff in 1937 and built in Iowa in 1942 with Berry's help to solve systems of linear equations. It had binary arithmetic and electronic switches, but was neither programmable, nor Turing universal. Unlike the ENIAC of 1947, which it inspired and which Wikipedia calls "the first electronic general-purpose digital computer". In 1973 a US court even invalidated the ENIAC patent, ruling that the idea of "electronic digital computer" was derived from Atanasoff.

Interestingly enough, "the world's first electronic analog computer" was built also in 1942 by Hölzer, an associate of Nazi rocket designer von Braun, to simulate V-2 rocket trajectories.

  • $\begingroup$ @RodrigodeAzevedo These are the designations assigned by Wikipedia, my point is that they are all based on fungible criteria. $\endgroup$ – Conifold May 1 '20 at 17:38
  • 1
    $\begingroup$ @RodrigodeAzevedo I do not really see such bragging rights as facts. If people want facts they should look at the timeline of machines and their features rather than draw imaginary lines on which was "first so and so", that is why I used scare quotes. $\endgroup$ – Conifold May 1 '20 at 20:11
  • $\begingroup$ That sounds wise. In any case, I convinced the OP to add "digital" to his question. I removed my previous comments, as they're now (twice) obsolete. $\endgroup$ – Rodrigo de Azevedo May 2 '20 at 3:35

Why does being decimal disqualify ENIAC? Decimal electronic stored-program computers used to be a thing, for example the commercially successful IBM 1401. It was a matter of "binary for scientific use, decimal for business use".

My own bias as a programmer disqualifies ENIAC for not being a stored-program digital computer, or more accurately, I'm more interested in stored-program systems. The honors for first place then go to the Manchester Baby. The EDVAC was designed before Baby, but for actually running, Baby wins. That's another bias: a working system.

As the other answer says, the question is ambiguous.

  • 1
    $\begingroup$ Thanks for the answer. Do you know why decimal was standard in bussines? Was it because business users were used to tabulating machines using decimal convention? $\endgroup$ – Martin Vesely Apr 13 '20 at 7:11
  • 4
    $\begingroup$ @MartinVesely see Why not use Double or Float to represent currency? $\endgroup$ – Ruslan Apr 13 '20 at 9:34
  • 2
    $\begingroup$ @MartinVesely, decimal was standard for business because the main thing businesses used computers for at the time was accounting. (These days, accounting software uses fixed-point math or custom math routines, either of which does a high-quality simulation of decimal computation.) $\endgroup$ – Mark Apr 13 '20 at 20:49
  • 1
    $\begingroup$ @MartinVesely Besides business, decimal was preferred for teaching purposes in those early computing days, because binary was considered confusing for beginning students. My early computing experience was a part-time job with an IBM 1620 (a decimal machine) in 1966, as an operator/programmer at NYU's School of Business. I was an undergrad in physics at CCNY at that time, so had no trouble at all with binary. But I helped many NYU business students with their programming homework, and they indeed frequently found binary confusing on the occasions where it cropped up. $\endgroup$ – John Forkosh Apr 15 '20 at 6:03
  • 1
    $\begingroup$ @MartinVesely Business computers read decimal numbers as input, did some minimal calculations (e.g. interest) and then printed decimal numbers as output. It was most efficient to stay decimal all the way. Scientific computers read decimal input, did a lot of calculations, then printed decimal output. It was most efficient to do the calculations in binary. $\endgroup$ – Stig Hemmer Apr 15 '20 at 7:22

My personal marker is stored program control, ie, the program is stored in the computer memory. Early machines like the original ENIAC, Colossus, Zuse's Z-machines, were all programmed by changing the wiring in some fashion. The Z-machines used relays, ENIAC was an advance because it used electronics (valves, the transistor hadn't been invented). The first stored program computer was the Manchester Baby in 1948.

  • 1
    $\begingroup$ "valves" (UK) = "(vacuum) tubes" (US) $\endgroup$ – Gerald Edgar Apr 18 '20 at 14:49

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.