Get best of the week

Dinosaur Walk: Two Large Computers of the 20th Century

In the famous Hollywood movie "Mind Games" there was such a curious scene: colleagues presented the pens to the Princeton professor as a tribute to his talent. It doesn’t matter if there is even a drop of reality in this beautiful metaphor, or whether it, like many other beautiful things, was invented on the conveyor of a dream factory. The ordinary pen (or rather, the reluctance to produce tons of mathematical calculations with it) was the inspiration for the first computers. Today we will talk about two of them.



Instead of a foreword


Quite a lot of articles were written about the first computers. Someone looked at them casually for granted. Someone went into details, described their internal structure, analyzed the historical prerequisites for their creation and their role in world technological progress.

Of course, all these aspects of the history of computers are extremely important and useful. They help us understand why the world in which we live today works the way it works.

However, this does not negate some of the boring and didactic nature of such materials. They leave the reader face to face with figures and facts and do not give a three-dimensional view of the computer.

As part of a small experiment, we would like to tell you about the two “first” computers: Aiken's machine, which he created in the 1940s together with IBM, and the Whirlwind project, which is not often mentioned in such articles, despite its notable fate.

First of all, we would like to show these computers “face”: to find interesting archival documents, bring photographs and video recordings. The detailed story of Mark I and Whirlwind has already been repeatedly told to us.

The first "computer" by Howard Aiken




Meet this man's name is Howard Aiken. It is generally believed that it is to him that we owe the appearance of the world's first device, which, with some stretch, can be called a programmable computer. The project he proposed was supposed to replace a person in solving differential equations that have only numerical solutions. Aiken came up with the idea of ​​creating such a device in the mid to late 1930s while working on his doctoral dissertation at Harvard. After a series of very successful attempts to assemble simple calculators suitable for solving a narrow range of problems, he thought about something more complete and interesting. By his own admission, the future car was inspired by the work of Charles Babbage.
I'm afraid if Babbage lived 75 years later, I would have been out of work
Howard Aiken

Babbage counting mechanism, photograph from the 1946 ASCC (Mark I) instruction manual. The full version of the manual can be found here .

Having made a detailed description of the concept of a future computer and secured financial support from the U.S. Navy (Aiken was a second-rank captain), he set about pacing the thresholds of companies that would help him put the project into practice: provide technical expertise and provide an adequate element base. But not everything was smooth. Nobody wanted to believe in Aiken’s car. Someone refused because of the excessive complexity of the project, someone because of the lack of market prospects.

Below is an excerpt from an oral interview with Robert Campbell, one of Aiken’s colleagues at IBM, as well as an abridged translation.
CAMPBELL: 1937 (, ASCC. — . ). It described functionally a machine which had a rather complete repertoire of characteristics, but said almost nothing about how it might be constructed or what components would be used. What circuitry techniques or what other mechanical techniques as far as that's concerned. It did, however, talk about tape programming I think.

/ 1937 , , . . , , ./

ASPRAY: What did that mean? / ?/

CAMPBELL: Well, a sequence control device from instructions somehow encoded in the punch paper (in a tape, not necessarily punched paper). But other than that there was very little dealing with specific components or specific design techniques. So having developed this concept Aiken tried to find a way of implementing it. He didn't have the resources himself. He was an instructor at Harvard at that point finishing up his doctorate degree. He first went to a number of business machine companies. It was at Monroe that he talked to George Chase, chief engineer. Chase was quite interested in the concept and would like to have tried implementing it using necessarily mechanical techniques, but the top management at Monroe was not interested.

/ , - . , , . , . . , . , . , , - ./

ASPRAY: They just didn't see a market for it, or what? / , ?/

CAMPBELL: Either they didn't want to spend the money for it or they didn't see a market for it either. He went to other business machine companies. I don't know how many. But he was unable to get any interest. Then through Professor Theodore Brown at the Business School and Harlow Shapley in the Astronomy Department he secured a introduction to Bryce — James Bryce of IBM, who was a senior executive in New York with a long history of engineering inventions. He became interested in Aiken's ideas; and through Bryce it was arranged for Aiken to talk to Watson. <…> In any case, Watson became quite interested in Aiken's idea.

/ , . . , [ ]. . Business School IBM, -. ; . <…> , ./


The further development of events is already more or less known to most readers: Watson, president of IBM, turned out to be a visionary man with a rich imagination and warmly supported Aiken's idea. We can say that Aiken pulled out a lucky ticket, starting a partnership with IBM: in those years, few companies had access to such a rich selection of technical components. In Aiken’s own words, if IBM had refused him, he would have to go to RCA or Bell Laboratories, and his machine could become completely electronic, rather than electronic-mechanical. Or not to be born at all. Summing up, it is worth noting once again: Aiken came to Watson with a project containing only a description of the requirements for the functionality of the machine. All technical implementation belongs to IBM and to some extent Aiken himself and his colleagues.At IBM, the project was named ASCC - Automatic Sequence Controlled Calculator. Thomas Watson was extremely proud that the first such machine would be released in his company, but at the end of the project his joy was overshadowed.

During the ceremonial handover of the finished test machine to Harvard, Aiken did not mention IBM's contribution to its creation in his speech. Thomas Watson became furious and broke off further cooperation with Aiken. The name “ASCC” was replaced by the paramilitary Mark I, and Aiken and IBM parted ways.

Let's make one more remark about Aiken’s personality. According to the same Campbell, Aiken belonged to the category of mathematical engineers. His real interest in computers lay purely on a practical plane. Computations were in the first place, and cars remained only a convenient auxiliary tool. He did not create a computer for the computer. Nevertheless, when in 1944 the monster for $ 200,000 (according to other sources, for $ 500,000) was finally completed and delivered to Harvard, Aiken with enthusiasm, no less than in development, started using it.

Here is a brief description of the resulting machine.

  • Over 765 thousand components.
  • About 17 meters in length.
  • 2.5 meters in height.
  • Weight: 4.5 tons.
  • Synchronization of modules using a 15-meter shaft with an electric motor.
  • At the insistence of Thomas Watson, the machine was enclosed in a “closet” of glass and stainless steel.
  • The machine could replace up to 20 people with manual devices for performing calculations.
  • The machine was programmable and did not require human intervention during operation.
  • Despite all of the above, the car was extremely slow even for its time.


A short story about Mark I and a demonstration of the appearance of the car these days


Mark I characteristics from the US Department of Defense reference

It is customary to say that Mark I was not really used in any serious calculations for 15 years of its work. Let us dwell on this moment in more detail.

Howard Aiken was appointed head of the Harvard Computing Laboratory, which, in turn, was at the disposal of the US Shipping Bureau. Of course, any technical advantage during the Second World War could be decisive, and the bureau of shipping was extremely positive about the appearance of a computer. Nevertheless, due to the specifics of the bureau, there were few orders for settlements with Mark I. Much more calculations were done for the ammunition bureau research laboratory. Work on the machine took place in three shifts, and in the “free” time it was used to calculate Bessel tables, in which Aiken saw its main purpose.


Fragment of the preface from the 1946 ASCC instruction manual (Mark I)

The next computer project, Mark II, was developed at the end of 1944 by Aiken and Campbell. It was planned to create another electromechanical calculator, this time for the needs of the Dalgren naval training ground. He was followed by two more similar cars, Mark III and Mark IV, respectively. Aiken's last computer no longer contained mechanical components and used memory on magnetic cores.

To date, Mark I is still operational. In 2014, it was serviced and launched.


POV Mark I


The same, from a different perspective.

The official shutdown of the “computer” took place in 1959 in connection with its complete obsolescence.

Whirlwind I Project: First Display Computer


Perhaps the Whirlwind project is one of the most interesting computers ever created. His brief history is as follows: in the early 1940s, the US Navy needed a simulator for pilots that would not be tied to a specific model of the aircraft and could be reprogrammed at any time immediately before the training session. The development was entrusted to a team of engineers led by Jay Forrester of MIT.



The catch was that the Navy needed not just a computer, but a system that responded to operator actions in real time. In the first year of development, Forrester focused his team on creating an analog machine capable of calculating the trajectory of the aircraft, but this solution was too complicated and not flexible enough for such a task.

In 1945, Forrester's group began to study the basics of digital technology and design a new computer using new technologies.

The technical complexity of the project (up to 100,000 operations per second was required against the usual 1000-10,000 at that time), the creators did not have such experience (the first fully electronic ENIAC computer just started working, a favorite of the girls, respectively, a new machine had to be developed with zero) and a host of other factors stretched the project over time. At the very early stages, the group leaders decided on the need to split the development of the project into two phases. At the first stage, it was planned to build a simplified machine, Whirlwind I, so that, taking into account its mistakes and the accumulation of experience, then create a more advanced, powerful and expensive computer.



The first detailed computer project was ready by 1947, after the war, and the end of testing all critical nodes was planned for 1949. However, history has made adjustments to this plan.

The high cost in the development of Mark I was partially offset by its military use. But at the end of World War II, the US government drastically reduced all military budgets, and the unfinished computer lost its main goal: the Air Force didn’t need to train pilots so intensively, there were enough “conservative” methods, which, moreover, were worth it significantly cheaper. According to average estimates, the development of Whirlwind, in view of attracting a significant number of both human and technical resources, cost the United States about $ 1.8 million per year (which in terms of modern money gives as much as about $ 18 million).


Whirlwind Floor Plan

Oddly enough, oddly enough, the EDVAC computer being developed in parallel was another stone towards the economic inexpediency of the Whirlwind project, the costs of which turned out to be significantly lower.

It is quite natural that the flight simulator project was minimized and reorganized into a purely “computing” one. The patronage of Whirlwind in 1948 passed from the Navy to the US Navy, which generally supported the development, but advocated for a significant reduction in price and simplification of the computer.

Forrester and Everett, development team leaders, turned to find a new customer for their project. They were made by the US Air Force, which, as part of the construction of an air defense system, needed a machine capable of combining data from multiple radars into a single map. By 1949-50, most of the finance came into the project from this department.

Whirlwind demo for television. See It now, 1951

Whirlwind I was commissioned in the early spring of 1951. As expected, the bulk of the machine time was given to the needs of the US Air Force. You can read more about using Whirlwind for military purposes here . Below is a scan of the project specification on the pages of the 1953 document .



Whirlwind I was also the first computer to receive a graphic display and a light gun as a control:






An amateur video made at the Smithsonian Institution Museum, where one of Whirlwind fragments is exhibited.

Immediately upon completion of construction, the computer was put into operation. With it, the US Air Force tested a new air defense system, the Cape Cod System. The project was curated by a specially created Lincoln Laboratory.

The tasks of the computer included solving problems of tracking targets for fighters. A special graphic display, which we wrote about a little higher, was created specifically for the convenience of operators: the system allowed you to "point" the light gun to the desired target. Another display, optional, was used to create copies of the screen - the camera was pointed at it. By 1953, the “Cape Cod System” coped with tracking 48 targets using 14 radars.

In addition to purely militaristic tasks, the computer was also used by scientists for calculations. The main requests came to the laboratory from ONR and MIT. According to some reports, a computer solved up to 200 such tasks per year.

The development of the Whirlwind project did not end on the creation of the first model. The second version of the computer, Whirlwind II, by 1959 surpassed its predecessor many times and the further use of Whirlwind I was considered economically inexpedient. The first shutdown of the computer took place in June 1959. However, the computer did not stand idle for a long time and was soon leased to Wolf Research and Development Company, founded by a former member of the Whirlwind project. For about 5 years, the computer worked in the office of the company, solving its tasks.

A complete shutdown of the computer took place in 1964. Whirlwind I was disassembled, and one of its fragments was transferred to the Museum of Computer History (video above).

To summarize a little in numbers:

  • computer weight: 9.1 tons;
  • project costs: $ 1 million per year;
  • approximate activity time: 14 years intermittently;
  • Whirlwind I introduced the world's first operator-driven graphical user interface.

We hope you found our review article interesting. In any case, looking at dinosaurs of computer technology while holding a mobile phone or sitting in front of personal computers is very funny and unusual.

In conclusion, we offer you several thematic links:

More articles:


All Articles