“With the advent of everyday use of elaborate calculations, speed has become paramount to such a high degree that there is no machine on the market today capable of satisfying the full demand of modern computational methods.” – part of the ENIAC patent (U.S.#3,120,606), filed on June 26, 1947, submitted by John W. Mauchly and J. Presper Eckert.
Mauchly and Eckert in front of their UNIVAC computer
John William Mauchly was born on August 30, 1907, in Cincinnati, Ohio. His father, Sebastian Jacob Mauchly, was a physicist who did teaching assignments in local high schools and colleges. When father Mauchly was offered a position as head of the Department of Terrestrial Magnetism at the Carnegie Institute in Washington, D.C. in 1913, the family, including John and his sister Helen Elizabeth, moved to Chevy Chase, Maryland – this is where John grew up. He attended the McKinley Technical High School in downtown Washington. He showed an early interest in science and engineering and loved tinkering with things and taking them apart. Because of his academic achievements in high school the State of Maryland awarded the young Mauchly a scholarship in 1925 which enabled him to study engineering at Johns Hopkins University’s School of Engineering. Two years later he didn’t find the subjects challenging enough and decided to switch to the department of physics. Recognizing his ability and academic interest, the department offered him a position in its Ph.D. programme. He completed his Ph.D. in 1932. This was the time of the Great Depression and work was difficult to come by, even for a graduate in physics who had specialized in molecular spectroscopy. During 1932-33 he stayed at Johns Hopkins as a research assistant. During this period he had to do many complex calculations which all had to be done manually – a time-consuming affair. This caused Mauchly to start thinking about the feasibility of a device to automate these mathematical calculations.
At Ursinus College
In 1933 Mauchly managed to get a position as professor at the one-man (himself) Physics Department of the Ursinus College, a small arts college near Philadelphia, Pennsylvania, where he lectured from 1933 to 1941. Apart from his lecturing duties he also managed to do research in meteorology which again required complex and lengthy calculations for which the manual calculators of that time were not really suitable. It revived his interest in automating computations and again he started considering an electronic machine – possibly using vacuum tubes – to accomplish complex calculations. Not being an expert in electronics and vacuum tubes Mauchly decided to learn as much as he could about these subjects. He also researched what other developments were taking place in the area of automating computations. In 1941 he visited John Atanasoff at the Iowa State university who was in the process of building his Atanasoff-Berry Computer (ABC). The ABC did not impress Mauchly as it was a special purpose device that could only solve linear equations while Mauchly was already thinking along the lines of a general-purpose computational device.
Mauchly and Eckert meet and team up
To further his knowledge in electronics Mauchly took a summer course at the Moore School of Electrical Engineering at the University of Pennsylvania in 1941. The course was primarily organized for army personnel to improve their knowledge of electronics as the army realized that the clouds of war were gathering. Mauchly didn’t find the course too difficult and impressed the Moore School staff to such an extent that he was offered a position as adjunct instructor in electrical engineering. During the course he also met one of the laboratory assistants, the brilliant undergraduate engineer J. Presper Eckert, whose task it was to assist the attendants with the lab equipment – or as Eckert himself later stated, “to make sure they didn’t break anything”.
Mauchly loved discussing his ideas about a possible calculating machine and found in Eckert a willing and enthusiastic listener. Eckert convinced Mauchly that vacuum tubes could be made reliable enough for this purpose.
The Moore School and the Ballistic Research Laboratory
With war imminent, new artillery had been developed and all trajectories for shells and bombs had to be calculated manually, depending on weight, propellant, distance, elevation, wind and similar variables. It was the task of the Ballistic Research Laboratory (BRL) at the Aberdeen Proving Ground in Maryland to publish these ballistics tables and perform the necessary tests of the artillery using these tables. For this purpose BRL had contracted the Moore School to do the calculations required to produce these tables. The complex and time-consuming computations were done by mostly female staff, all college graduates, who were called “computers”. Interestingly, Mauchly’s second wife, Kathleen McNulty, was a former “computer” who later became a programmer on the ENIAC. It took one skilled “computer”, using a mechanical desk-calculator, about 20 hours to calculate a ballistic trajectory. The U.S. Army and Navy had a desperate need for these firing and bombing tables and the Moore School was under permanent pressure to produce the ballistics tables faster. It made the Moore School the center for war-time computing.
This photo of 1946 is featured in the margin of a souvenir sheet issued in 2000 by the
Central African Republic. The originally black-and-white photo shows the ENIAC with
one of its maintenance engineers, Corporal Irwin Goldstein of the BRL, in the foreground.
In the background on the right is ENIAC programmer Jean Jennings (now Jean Bartik).
Mauchly and Eckert had many discussions on how to construct a general-purpose calculating device. Should the machine be partly electronic – partly mechanical, or fully electronic? Eckert recalled in a 1988 interview: “We finally came to the conclusion that if you're going to do this, you ought to do it whole hog and make everything in sight digital.” In 1942 Mauchly wrote a memorandum, “The Use of High Speed Vacuum Tube Devices for Calculating”, proposing to construct a general-purpose electronic computer. He emphasized that by using electronics instead of mechanical parts, the speed would be significantly faster than anything else available. In the same paper Mauchly also was the first one to use the verb “to program”. It took a while before the significance of his memo was understood but in 1942 Lieutenant Herman H. Goldstine of the BRL, a Ph.D. Mathematics, was assigned to the Moore School to supervise their computational and training facilities. When he heard about Mauchly’s memo, he asked Mauchly and Eckert to write a formal proposal to the BRL for this computing device. In June 1943 the first contract was signed for “research and development of an electronic numeral integrator and computer” with funding by BRL. The top-secret project was code-named Project PX. Nine supplementary contracts extended the work on the ENIAC – Electronic Numerator, Integrator And Computer, with a few sources listing the word “Analyzer” for the letter A – to 1946 at a total cost of US$486,804. The machine was to be built at the Moore School, followed by delivery to BRL at the Aberdeen Proving Ground. The project started with J. Presper Eckert as its chief engineer and John W. Mauchly as its principal consultant being responsible for its conceptual design.
1999 Dominica stamp – with the text ““Eniac” becomes world’s first computer”
– shows a 1946 photo of the ENIAC. The people shown on this stamp are (left to
right): Pfc (private first class) Homer Spence, electrical engineer from BRL;
J. Presper Eckert (in front), chief engineer; John W. Mauchly (standing behind
Eckert, carrying a cup of coffee), consulting engineer; Jean Bartik (née Jennings),
ENIAC programmer; Capt. Herman H. Goldstine, liaison officer BRL.
Jean Bartik started her career as a “computer” at the Moore School before being selected to become one of the six original programmers – all female – for the ENIAC. Bartik was the leader of the team to convert ENIAC to a stored-program computer after it was installed at the Aberdeen Proving Ground. In June 2009, at age 84, Jean Bartik received the IEEE Computer Society’s Computer Pioneer Award. She passed away on March 23, 2011, the last one of the original six ENIAC programmers.
The ENIAC was completed in 1946. It was a computer of behemoth proportions measuring 2.6m high by 0.9m deep by 26m wide. It occupied 63 m² of floor space and had a weight of 30 tons. ENIAC contained 17,468 vacuum tubes (an invention which via Ambrose Fleming and Lee de Forest can be traced back to inventor Thomas Edison), 70,000 resistors, 10,000 capacitors, 3,000 switches, 1,500 relays and 500,000 hand-soldered connections. By never switching off the machine and keeping current and voltage within strict limits as specified by Eckert, the failure rate for its vacuum tubes averaged to one tube failure every two days. It used an IBM card reader as input mechanism, an IBM printer as printing device and an IBM card punch machine for output.
The ENIAC had a capacity of storing twenty 10-digit decimal numbers – in contrast with modern computers which operate on binary numbers, the ENIAC operated on decimal numbers. It could do 5,000 additions per second, a division of a 10-digit number took 25 milliseconds while it did about 350 multiplications per second. Where it took the human “computer” 20 hours to compute a ballistic trajectory, the ENIAC performed this task in 30 seconds – an enormous improvement. ENIAC was the fastest computing device of its time made possible by Mauchly’s vision and his ability to implement all the different concepts, combined with Eckert’s superb ability as engineer and his painstaking attention to all technical details. To put things in perspective: today a silicon chip measuring 0.01 inches can have the same capacity as the ENIAC.
The ENIAC had a number of novel developments: it used subroutines for repetitive tasks which was unheard of till then. Another interesting feature was its ability to execute one addition at a time at the push of a button. This assisted greatly in the debugging of its programs. The author can testify to the value of such a facility as twenty years later the IBM 1401 had a similar facility for debugging programs.
Programming the ENIAC was done externally through wiring a plug board combined with the manually setting of switches. It was soon realised that this was a severe weakness as preparing a new program became a time consuming affair. To remedy this, Mauchly and Eckert, with the assistance of others like John von Neumann, already had started designing the next computer, the EDVAC (Electronic Discrete Variable Automatic Computer) in 1944 and had decided to implement a stored-program concept in the EDVAC – i.e. the program controlling the computer was kept within the computer’s own memory. Von Neumann wrote a report “First Draft of a Report on the EDVAC” in June 1945 for internal distribution, but Herman Goldstine controversially removed the names of all participants in the report and distributed it widely with only Von Neumann’s name on it. As a consequence acrimony arose within the team and history has since called the stored-program concept the Von Neumann architecture. It is used today in all modern computers.
The ENIAC became operational at the Moore School in February 1946. Although used for ballistics, one of its first uses was for calculations for the development of nuclear weapons. During 1947 the ENIAC was dismantled and transported to the Aberdeen Proving Ground where it became operational again in August 1947. There the ENIAC underwent a number of upgrades and modifications. It was used for ballistics, weather forecasting, atomic energy calculations, wind-tunnel design and other scientific uses. When its power was finally switched off at 11:45 p.m. on October 2, 1955, the ENIAC had been in operation successfully for a total of 80,223 hours. A remarkable achievement for the world’s first general-purpose electronic computer.
The U.S. Postal Service commemorated the birth of the ENIAC by issuing a “Computer Technology” stamp
fifty years later (1996). Appropriately, the launch of the stamp took place on the
Aberdeen Proving Ground and the event was broadcasted live over the Internet.
The Moore School Lectures
Because the Moore School was in the center of computer development it organized a course called “The Theory and Techniques for Design of Digital Computers”. The course ran for eight weeks from July 8 until August 31, 1946, and consisted of a series of 48 lectures. The course became widely known as the Moore School Lectures. The majority of the lectures was given by Eckert, followed by Mauchly and Goldstine. Because this was the first computer course offered anywhere, it generated enormous interest. Attendants came from universities, army and navy, but also commercial companies like IBM, General Electric, Bell Labs and NCR. The Moore School Lectures gave an enormous impetus to an industry that was still in its infancy.
The U.S. Census Bureau which was to use the first Eckert-Mauchly UNIVAC
was using Hollerith punch card technology (see enlarged indicium)
The Eckert-Mauchly Computer Corporation
The Technical Camelot
Poem by Jean Bartik – ENIAC programmer,
describing work at the Eckert-Mauchly Computer Company
There once was a congenial spot
Where ideas flowed so freely
And designs were done so speed’ly
Where all of them forgot
Frontiers weren’t pierced so easily
No one believed in 'Not'
In Technical Camelot
In 1946 the University of Pennsylvania instituted a patent policy that all patents pertaining to research done by faculty members belonged to the university. Since work on the ENIAC had been funded by the military and not by university funds, Eckert and Mauchly refused to comply and turn over their patent rights. In order to hold on to their patents they had to resign on March 31, 1946. That same year the two colleagues formed the Electronic Controls Company with Eckert in charge of designing a new computer system and Mauchly responsible for its programming and coding as well as marketing and sales, including the identification of new applications. Mauchly approached the U.S. Census Bureau as the U.S. national census was due in 1950 giving them four years to develop, design and build a new computer. As the U.S. Census Bureau was not permitted to fund research, partial payment up-front was done through the National Bureau of Standards.
In 1947 the company was renamed as the Eckert-Mauchly Computer Corporation. The computer to be built for the National Bureau of Standards and the Census Bureau was the EDVAC II which was later renamed to UNIVAC (UNIversal Automatic Computer). In order to raise sufficient cash to keep the company going, an order was obtained from the Northrop Aircraft Company for the Binary Automatic Computer (BINAC). It was a small computer for the calculation of air missile trajectories and was delivered in 1950. Since the BINAC stored its data on magnetic tape instead of punched cards, it operated significantly faster than the ENIAC.
Despite all efforts, the company continued to experience financial difficulties and was finally taken over by Remington Rand in 1950 where it became a separate division: the UNIVAC Division within Remington Rand. Both Mauchly and Eckert stayed on.
To commemorate the 50th anniversary of the computer Bosnia & Hercegovina issued
in 2001 (five years too late) a sheet with 13 se-tenant pairs and a four-stamp label
showing a photo of the UNIVAC. The photo was taken by John W. Mauchly.
Both stamps show strings of bits (“0” and “1”s).
In 1952 Remington Rand approached CBS and told them that their new UNIVAC computer might be able to process the early election results of the U.S. presidential election between Dwight Eisenhower (Rep.) and Adlai Stevenson (Dem.). The computer should then be able to predict the winner at a very early stage. Since the polls predicted a close contest and computers were a rather unknown novelty, CBS executives and its anchor Walter Cronkite had little faith in Remington Rand’s offer. Despite their misgivings they agreed to the plan as they thought it might be entertaining for the TV audience.
On election night – November 4, 1952 – CBS set up a camera crew in front of the UNIVAC which was located in Philadelphia, while in the studio Cronkite sat next to a mock-up computer CBS had built, consisting of a panel decked out in blinking Christmas lights next to a teletype machine. Using Unityper machines, clerks typed in the election results as they came in. The Unitypers produced a punched paper tape which was fed into the UNIVAC. At 20.30 E.T. the UNIVAC produced its first prediction: Eisenhower would win with 438 electoral votes against Stevenson’s 83 – a landslide victory! Considering earlier polls which had predicted a tight race, CBS refused to put the UNIVAC prediction on air, but said that Eisenhower might win with a small majority. Mauchly was contacted and the algorithm was re-checked and re-run. Again: Eisenhower would win by a mile. Later that night, as more actual results came in, CBS realized that the UNIVAC had been right. CBS went on air admitting its mistake and informed the world that the UNIVAC had predicted the landslide as early as 20.30 that evening. The final results were: Eisenhower – 442 electoral votes versus Stevenson – 89 votes. The computer had been out by 1%. UNIVAC’s name was made and the general public’s imagination had been captured by a computer's miraculous performance. It had been an extremely risky but highly successful public relations exercise.
Take-overs and mergers
In 1955 Remington Rand merged with Sperry Corp. to form Sperry Rand Corp. with Mauchly becoming director of its UNIVAC Applications Research Center. In 1959, when he was asked to change from research to marketing, Mauchly resigned from Sperry Rand and formed his own company – Mauchly Associates, Inc., concentrating on consulting and computer development. Here Mauchly developed a task scheduling and planning tool, called critical path method (now called CPM). In 1967 he set up his own consulting company, Dynatrend, which he ran until his death.
Eckert became an executive at Sperry Rand and decided to stay on when Sperry Rand merged with Burroughs to form Unisys in 1986. Eckert retired from Unisys in 1989.
John W. Mauchly died during heart surgery on January 8, 1980, in Abington, Pennsylvania.
John W. Mauchly received numerous awards and honors. Some awards he received jointly with J. Presper Eckert (for details see my article about Eckert):
In 1949 Eckert and Mauchly received the Howard N. Potts Medal from the Franklin Institute in Philadelphia.
In 1961 they both were awarded the John Scott Medal.
In 1966 they both were awarded the Harry M. Goode Memorial Award by the American Federation of Information Processing Societies.
In 1980 the Institute of Electrical and Electronics Engineers (IEEE) awarded both Eckert and Mauchly the IEEE Computer Society’s Computer Pioneer Award.
In 1957 Mauchly was elected Fellow of the IRE, the forerunner of the IEEE. He also was a life member of the National Academy of Engineering and the Society for Advancement of Management. He was elected Fellow of the American Statistical Organization. He received honorary doctorates from the University of Pennsylvania and from Ursinus College. He also received the Philadelphia Award and the Emanuel R. Pione Award.
Mauchly and Eckert are jointly portrayed on a stamp issued by Madagascar in 1993. In addition they are both featured on the 1999 Dominica stamp shown above. The U.S. has not issued a stamp to commemorate either of them.
© Wobbe Vegter, 2009