15

I was watching the DVD movie Imitation Game, starring Benedict Cumberbatch as Alan Turing, the British mathematician who helped crack the Nazi's enigma code in WWII. In one key scene, Turing uses the expression digital computer to explain to Joan Clarke what ‘Christopher’, the machine he is making, does. He describes it as

“... an electrical brain, a digital computer.”

I think the year depicted in the movie must have been 1940. The name of Turing's creation was never called ‘Christopher’ that is pure dramatic/artistic license, its real name was bombe and it was a monumental deciphering machine.

Basically, I have three questions:

  1. Was Alan Turing aware of the term digital computer? Did he actually ever use it himself?

  2. Could one describe the British Bombe as being a digital computer? Wikipedia calls it an “electromechanical device”

  3. When was the term digital computer actually coined?

‘Christopher’

YouTube clip: Alan Turing explains 'Christopher'


BOUNTY INFO

I should have left a message with the bounty, too late now.

Q1. I am asking whether Turing was familiar with the term “digital computer” in the 1940s but especially before the end of World War II. Biscuit Boy's answer refers to Turing's paper, Computing Machinery and Intelligence, which is dated 1950.

Q2. Because I know nothing about the history of computers nor how they work, I am hoping the community of Stack Exchange can provide a clear-cut answer.

Q3. According to this source, the term digital computer was devised by George Stibitz in 1942, which makes Cumberbatch's line a lexical anachronism.

Q4 Bounty bonus. The name of the first electronic digital computer.

Many cite ENIAC as being the holder of this title, but there are sources which claim the Universal Turing Machine “is the mathematical tool equivalent to a digital computer", and elsewhere: “The first fully functioning electronic digital computer was Colossus (1943)”.

25
  • 1
    Re: Q2. The Bombes were electromechanical devices that used relays (telephone exchange items) & rotors, but not electronic components such as tubes/valves (US/ British terms). The first generation of digital computers used tubes/valves (late 1940s into the 1950s). The second generation used transistors (1950s) & the third generation used integrated circuits (1960s). Integrated circuits have since developed into what we now call computer chips.
    – user122572
    Jan 22, 2016 at 13:30
  • @Fred - While you are correct about the nature of the Bombe, it was still technically a "digital computer". The real question is whether it ever occurred to anyone to call it such.
    – Hot Licks
    Jan 22, 2016 at 13:36
  • The term to use for what we now call computers was up in the air for a number of years in the 40s and into the 50s. As someone else noted, the main meaning of "computer" was a person who does computations. And earliest surviving professional organization dealing with these devices is, to this day, called the "Association for Computing Machinery", since "computer" was not a settled term when it was founded.
    – Hot Licks
    Jan 22, 2016 at 13:44
  • 2
    Thank you for posting this. This is an example of the kind of discussion that originally caught my attention on this site! :)
    – Tim Ward
    Jan 22, 2016 at 18:00
  • 1
    More on Colossus: news.bbc.co.uk/local/threecounties/hi/people_and_places/history/…
    – phoog
    Jan 22, 2016 at 19:04

6 Answers 6

8

I was going through some online articles and I'd like to thank @Josh61 for the right references. I found this detailed write-up on Word Origins from OED by Richard Holden. (I think I now know the reason why top EL&U users strictly stick to OED definitions)

The article: http://public.oed.com/aspects-of-english/word-stories/digital/

What distinguishes digital from many other terms associated with high technology is that it’s not a new word. In the newly revised OED entry, the earliest evidence—in the sense ‘designating a whole number less than ten’—dates from the fifteenth century. OED‘s original entry, published in 1897, does not record this sense. Instead, it covers senses corresponding to another sense of digit, such as: ‘of or pertaining to a finger, or to the fingers or digits’evidence for which goes back to the seventeenth century. But for most of its history, digital was a relatively unimportant term: it wasn’t until the early to mid-twentieth century that the word became more significant and widespread.

In the late 1930s and early 1940s, the work of mathematicians and engineers led to the development of a new type of computing machine. As opposed to earlier analogue devices, which used a continuous quantity (such as voltage) to compute the desired quantity by analogy, these new machines operated upon data that was represented as a series of discrete digits. For example, in such a system the letter A might be represented as the binary sequence ‘01000001’ (as it is in the ASCII encoding scheme).

Being composed of such sequences of digits, such data (and so any machine making use of it) was hence said to be "digital". Digital computers were generally considered more adaptable and powerful than their analogue counterparts, and digital computing became dominant: the computer you are reading this article on will certainly be a digital one, as will probably any other computer you have ever used. The sense of digital relating to this was covered in OED2 (1989) by the definition, ‘of, pertaining to, or using digits; spec. applied to a computer which operates on data in the form of digital or similar discrete elements.’


Now, to answer your questions...

1. Was Alan Turing aware of the term digital computer? Did he actually ever use it himself?

Yes. Alan Turing developed the Turing test, and uses the term often in his seminal paper "Computer Machinery and Intelligence".

"Are there imaginable digital computers which would do well in the imitation game?

[Wikipedia]


2. Could one describe the British Bombe as being a digital computer? Wikipedia calls it an “electromechanical device”

I would say No. At least not a "digital computer". It's called electromechanical because it operated on electricity and had rotors, wires and plugboards that carried out the deciphering work. Importantly it did not operate on 0s and 1s, which is the definition of a "digital computer".

(noun). an electronic computer in which the input is discrete rather than continuous, consisting of combinations of numbers, letters, and other characters written in an appropriate programming language and represented internally in binary notation

[Collins Dictionary]

The ENIAC is widely believed to be the first ever "digital" computing device.

It was Turing-complete, digital, and could solve "a large class of numerical problems" through reprogramming

[Wikipedia]

You could probably say it was an "analog computer"!

(noun) a mechanical, electrical, or electronic computer that performs arithmetical operations by using some variable physical quantity, such as mechanical movement or voltage, to represent numbers

[Collins Dictionary]


3. When was the term digital computer actually coined?

The exact time and person who coined it remains a mystery. Etymonline suggests that both "digital" and "computer" originated in the middle of 16th century but the combined term, i.e., "a digital computer" might have its origins at the beginning of the 20th century, and a spike in usage by late 1940s, according to Ngrams.

20
  • 2
    Well played and answered, Sir!
    – Mari-Lou A
    Jan 22, 2016 at 13:13
  • 7
    @BiscuitBoy Well, but it had all these rotors which, for instance, could take 26 positions. That is not binary, true, but still discrete, or saying it in another way, the number of all possible states of the machine is finite. Jan 22, 2016 at 13:28
  • 9
    Operating on ones and zeros is not the definition of "digital". A digital computer operates on digits, vs being "analog", and the Bombe and several other mechanical units of that era did, to my knowledge, operate mostly on digits. Most early electronic computers operated in decimal, or some variant thereof.
    – Hot Licks
    Jan 22, 2016 at 13:40
  • 5
    Agree with @HotLicks. Digital is not synonymous with binary. Binary is a subset of digital. Babbage's Analytical Engine that he described in 1837 was a digital computer.
    – ghoppe
    Jan 22, 2016 at 15:58
  • 3
    Turing's Bombe was most certainly not an "analog computer". From Wikipedia: "An analog computer is a form of computer that uses the continuously changeable aspects of physical phenomena such as electrical, mechanical, or hydraulic quantities to model the problem being solved." From a 1959 paper: "Operation of the analog computer is on a continuous basis, with the electrical parameters behaving precisely as the continuous physical system or equation variables do." There was nothing continuous about the Bombe. Feb 2, 2016 at 17:49
3
+100

In Computable Numbers (1936), Turing doesn't use the word "digital" at all. He refers to "computing machines" and in particular the "universal computing machine", which is what we now call a Turing machine. In modern terms, we would not usually refer to something as a computer unless it represented a concrete implementation of a Turing machine (except that strictly speaking, a Turing machine has infinite memory "tape"). The Turing machine is an abstract concept, and physical implementations of it don't necessarily have to be "digital" in any particular sense, but that's a complicated subject, and anyway all modern computers are digital.

Obviously, Turing didn't say much publicly during the war, so we don't know if he was using the term "digital computer" or not. In 1945 he moved to the National Physical Laboratory to develop the Automatic Computing Engine (ACE), which was his choice of name in homage to Babbage's machines[1]. In [Copeland BJ ed., The Essential Turing Oxford 2004], a 1947 lecture on ACE is reprinted without further bibliographic detail; Turing describes ACE as an "electronic digital computing machine" and goes on to emphasise the significance of the word "digital". He also refers to his earlier work on computability, saying "I was investigating what might now be described as... digital computing machines" (my emphasis). This suggests to me that he wouldn't have used the term in 1936, but he was by 1947. His discussion of the advantages of digital computing is essentially to do with the practical implementation of what was a purely theoretical thing in Computable Numbers, so a reasonable guess would be that he started using the word when he became involved in actually building machines; so, certainly by 1946 and probably in conversations related to the Mark 1 (Colossus) in 1943. In any case, I don't think the word would have been unfamiliar to him at any time during the war; there's a good chance he would have been specifically aware of Atanasoff's work, for example.

The thing that struck me more about that Imitation Game line was not the word "digital" but the word "brain". Turing thought a great deal about the essential nature of thinking, and I don't know whether he'd have casually conflated "brain" and "mind" in that way. It'd be an interesting comment, if he'd actually said it.

The first Turing-complete computer to be designed was Babbage's Analytical Engine in 1842, and the first to actually be built was ENIAC in 1946. The first machine with a von Neumann architecture-- very close to what we think of as a computer now-- was EDVAC in 1949. ACE wasn't completed until 1950.

The problem with counting "Colossus" as a computer is that it wasn't Turing-complete, or particularly general-purpose, so it's then hard to justify withholding the title from much older machines, especially the German Zuse Z3 from 1941. In fact the Z3 has been shown to be technically Turing-complete, but more as a theoretical stunt than in any practical sense. 19th-century mechanical calculators could do calculations, for example, and Jacquard looms were programmable.

The Polish bombes, which were later adopted and improved by Bletchley Park, were not computers by most definitions, either now or then. They weren't programmable, and didn't really perform any calculation except that they'd stop when they hit a certain combination of electrical settings; if we consider that computation, then many other machines, such as door locks, would qualify.

1
  • Because you answered the bounty questions pretty magnificently to a layperson such as myself.
    – Mari-Lou A
    Feb 8, 2016 at 11:00
2

It boils down to what you mean by "digital" (and, to a degree, "analog"). The best-known dichotomy in the computer universe is between "digital" and "analog". An "analog computer" is one which represents values in a computation as non-discrete voltage or current levels within an electronic circuit. A "digital computer", on the other hand, represents values as discrete voltage or current levels (or, for that matter, something such as the position of a "stepping relay" in a mechanical device).

Of course, one can do all sorts of dancing around both terms. One could argue, eg, that since the Bombe contained within it a simulation of (multiple) Enigma devices, it was an "analog" computer. However, one could then make the same claim of an Macintosh that is programmed to simulate Enigma.

What most people don't understand is the meaning of "digital". In particular it does not mean "binary" -- there were many computers built before "binary" became the default mode of operation. Most (especially the the mechanical ones) were decimal, but a number used other bases. "Centesimal" (base 100) was fairly popular, and a few others used ordinary Roman characters (generally with the "smarts" to do decimal arithmetic on numeric characters). "Bi-quinary" (a rather confusing sort of two-part decimal, probably based on the abacus) was quite popular.

Even in purportedly "binary" computers, sometimes multiple bits of data are represented in a single electrical signal by using multiple discrete voltage levels. (I don't know about now, but ca 1980 a large proportion of memory chips used this technique.)

So the main difference between "digital" and "analog" is that "digital" uses discrete voltage/current levels to represent values (and hence, if arithmetic is to be performed, the arithmetic is performed in discrete units called "digits", though those "digits" many be 1 bit, a decimal value, a centesimal value, or something else). Analog, on the other hand, performs simple addition by using a resistor network or some such to "sum" voltages or currents -- electrically adding the values together the same way an audio mixer "adds" two microphone signals together.

And, while it's not entirely clear without further study, it appears that the Bombe did actual digital arithmetic of a sort, as the encoding algorithm apparently (based on my vague understanding) involved "rotating" character values by some amount (incrementing an "A" to a "D", say) to encrypt them. This would have required a crude digital adder.

The harder question is whether the Bombe classifies as a "computer", in any modern sense. This is tricky, since we now consider a "computer" to include writable program store, but that was not the case of many early devices which were referred to as "computers".

4
  • However, the coinage "digital computer" was a modification of the word "computer," which described a person whose role was to perform computations. Feb 2, 2016 at 17:38
  • @outisnihil -- Actually, "computer" was the modification of the word "computer". Then they started getting more specific. The oldest professional organization of computer geeks is the Association for Computing Machinery, since the term "computer" did not have a settled meaning when the organization was formed.
    – Hot Licks
    Feb 2, 2016 at 17:54
  • I believe that the way it worked was that "digital computer" was coined in opposition to "computer" (the person who did computations), and then eventually "digital computer" was shortened back to "computer." The citation for "digital computer" is from 1942, while the ACM was founded in 1947. But I'm happy to be corrected on that! Feb 4, 2016 at 17:20
  • 1
    @outisnihil - Don't forget that there were also "analog computers" in that timeframe. The Norden bombsite was, after all, an analog computer.
    – Hot Licks
    Feb 4, 2016 at 21:37
2

The ENIAC was the first major electronic general purpose computer of major significance, using valves as its computation and storage devices. That made it orders of magnitude faster than its electromechanical predecessors. It was completed in 1946.

An seminal earlier electromechanical computer was the Harvard Mark I, completed in 1944 but started quite earlier.

The British Colossus computers were used as early as 1944. While not general purpose (programming was through plugs and switches), they also were electronic and digital.

With regard to electromechanical computers, Zuse's Z1 and its successors may be the earliest serious contenders. Most certainly Zuse was the first to fully employ binary numbers (rather than decimal numbers, typically encoded in BCD or excess-3 BCD).

It took a long time before binary numbers were used to the exclusion of everything else: even Knuth's seminal "The Art of Computer Programming" series first published in 1968 specifies a hypothetical computer architecture that may be either binary or decimal.

So all of "digital", "electronic", "binary", "general-purpose", "programmable in memory" have had different origin stories and dates.

1
  • The second link wasn't working, for some reason, it's now fixed. I hope the other links from Wikipedia were the ones you wanted.
    – Mari-Lou A
    Feb 1, 2016 at 17:25
1

Two rather different projects that Alan Turing worked on appear to be involved.

  1. Turing was a key person in breaking the German ciphers during WW2. (Many other people/groups were involved also.) To that end, many devices were constructed that mimicked the German ciphering machines, using parts of what the British were able to scrap from German submarines, etc. After many tedious and failed attempts, and sometimes only intermittently, the British were thus able to decipher the communication of the German naval forces. This is unrelated to digital computer.

  2. Turing defined a computation model, nowadays referred to as the Universal Turing machine. Building upon his observations on how humans compute by hand, he lay down principles of step-by-step computation, including key concepts such as memory (the tape of the machine) and instruction set. Again, many other people were involved (von Neumann, to name just one). Turing's ideas on computation are given is his seminal 1936 paper On Computable Numbers, with an Application to the Entscheidungsproblem.

The universal machine, mentioned in the clip, is defined in the above paper, and served as a model for computers as known nowadays. It is not a gross exaggeration to say that Turing invented the computer. Notably, Turing realized that the operational instructions of a machine, can be read and processed by another machine; thus, a single machine can be programmed to process infinitely many tasks rather than just one.

I do not know whether the collocation "digital computer" would have been meaningful to Turing. He uses the term machine for the concept that later yielded the computer of today. I think Turing viewed computation as inherently digital; so he might have viewed the qualification digital as somewhat redundant. As such, it is just a buzzword in the movie. (That said, there is no reason to think the creators of the movie did not know exactly what they were doing; they certainly did not want to explain the Turing machine in Turing's own terms! And it appeared not much later; John von Neumann certainly used it by 1945.)

It is to be noted that computation has a distinguished history in the British Isles; one of the most notable predecessors is the Analytical Engine, of Charles Babbage and Ada Lovelace (mechanical, of course). In that light, a quirky point about the Turing machine is that it is in fact not a machine, but rather an idea of it; it is a design, a model.

9
  • OK, I accept that the film makers used to great effect the expression "digital computer", in fact it is only used once in the entire film. But today, would Turing's "bombe" be considered a digital computer, albeit a very rudimentary one?
    – Mari-Lou A
    Jan 22, 2016 at 18:43
  • @Mari-LouA as discussed in other comments, one can argue that it is, but it's a bit of a stretch, and the real innovation being discussed in the clip is programmability, being what computer scientists call "Turing complete." Such a machine is a general-purpose machine; the bombe does not have this property. It is a single-purpose machine. There are huge liberties being taken, not only with regard to this plot point, but also with respect to Turing's role in Bletchley park.
    – phoog
    Jan 22, 2016 at 19:12
  • If I remember the movie correctly, the plot also mixes up some aspects of the Colossus project with Turing's work. Colossus was a reprogrammable digital computer.
    – phoog
    Jan 22, 2016 at 19:15
  • @Mari-LouA No, I don't think so. I do not understand all the details, but Bombe appears to be a one-purpose device, designed to speed up the process of searching for the right configuration, which is impossible to do by hand. I do not think it contributes anything but speedup by many orders, so it makes an impossible task possible (just so). But it is not programmable, as phoog says, which is ultimately the brilliant idea of the universal machine. On the other hand, one can argue that everything that a computer can do ...
    – anemone
    Jan 22, 2016 at 19:20
  • 1
    @anemone Sorry, that was in response to Mari-LouA's comment. Feb 4, 2016 at 17:16
1

Digital computation dates back much further than you think, to the origin of numerals them selves; It is of interest that essentially, digital computation can be assimilated to two counting systems; that of base ten or decimal, in which we can do math whilst counting upon our fingers, using both hands. Secondly, to a base twelve system, the dozen; a method in which one can use the thumb to count upon the phalanges of one hand, surprisingly somewhat more efficient for keeping an account.

It is note worthy to remark upon the use of the fingers, and hand gestures, when measuring the degree of separation between the asters in the night sky. This is quite naturally considered as the origin of the digital counting machine. However history is so often distorted by those who decimate; That origins become aberrant, such are the dipping waters of time ...

Let us not forget that Alan Turing was solving Cypher/Cipher and his machine was replacing his fingers and phalanges with cogs cams and wheels.

I do not profess any definitive answer to your question; I do however hope that you might find food for thought, upon your quest for knowledge itself; The origin of digital computation.

3
  • Yep, the Sickly Green Marchant was arguably a digital computer.
    – Hot Licks
    Feb 2, 2016 at 17:58
  • (Anyone hoping to understand what a computer is must first read Also Sprach von Neumann. Unfortunately, I don't believe that the entire thing (4-5 installments) is online on a single site, and it may be necessary to dig up old Datamation issues in a college library to get all of it.)
    – Hot Licks
    Feb 2, 2016 at 18:06
  • Yes, you can use digital and counting as synonyms: “... an electrical brain, a [counting] computer.” Feb 3, 2016 at 14:44

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

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