Feeling a little geeky? Nostalgic for the days when NASA had less computing power than your cell phone? In honor of Women's History Month, the Libraries and the National Museum of American History would like to enable you to build your own ENIAC. Ok, well, maybe not really build – you might have trouble finding over 17,000 vacuum tubes – but you could learn how to run one!
First a little context…we were thrilled a few months back when Peggy Kidwell, curator of Mathematics at the National Museum of American History, decided to transfer a piece of printed computing history to our Dibner Library of the History of Science and Technology. It was a complete set of the operational and technical manuals for the first 'general purpose' computer, ENIAC.
Left: Betty Jennings (Mrs. Bartik) Right: Frances Bilas (Mrs. Spence) operating the ENIAC's main control panel while the machine was still located at the Moore School. U.S. Army Photo from the archives of the ARL Technical Library. [Image from Wikimedia Commons ]
ENIAC, short for Electronic Numerical Integrator And Computer, was built at the Moore School of Engineering, University of Pennsylvania for the U.S. Army Ballistics Research Laboratory and completed in 1946 after the end of World War II. ENIAC was designed by J. Prespert Eckert and John Mauchley, created by many talented engineers and mathematicians at the Moore School, and programmed, debugged and operated by 6 women - Kathleen McNulty Mauchly Antonelli, Frances Snyder Holberton, Marlyn Wescoff Meltzer, Frances Bilas Spence, Ruth Lichterman Teitelbaum, and Jean Jennings Bartik.
ENIAC grew out of a need to more quickly and efficiently perform the computations to create ballistics trajectory and firing tables. As the majority of men on college campuses had been drafted for the war, the computations required for creating those tables were primarily done by women (many of whom were themselves mathematicians or math majors) operating simple mechanical calculators. As ENIAC was being developed, some of the same women who had been employed as 'computers' for the Army were then hired as 'operators' for the new electronic computer. Their jobs turned out to be much more than just 'operating' the computer. For starters, when the original programmers were hired in 1946, because the computer itself was not yet finished and no manuals had been written they had to learn how ENIAC worked by studying circuit diagrams.
Programming ENIAC was a complicated and time consuming process that involved figuring out the logical sequence of operations that would correctly perform the calculations; manually moving and plugging in cables; physically flipping switches; creating the input and interpreting the output (both on punch cards); and troubleshooting the 40 parallel units that made up ENIAC with 18,000 vacuum tubes,1,500 relays, 70,000 resistors, and 10,000 capacitors. Programming to do one calculation often took days and could take longer if say, cleaning staff accidentally knocked out a cable and plugged it back in to the wrong plug! Eventually, in 1948 a simple read-only storage mechanism was devised that greatly reduced the time it took to program ENIAC. This new stored-program process was developed by a team of programmers and engineers, including four programmers lead by Jean Bartik, Richard Clippinger,John von Neumann, Adele Goldstine, John Giese, and A. Galbraith. All 6 of the original ENIAC programmers were inducted into the Women In Technology International Hall of Fame in 1997.
Adele Goldstine, one of the original 6 programmers, also wrote the technical manual and the Report on the Eniac, copies of which are now at the Dibner Library of Science and Technology. The Libraries is happy to report that we are now digitizing all the operational and technical manuals, including diagrams, and will be making them available online at the Internet Archive. We have finished Report vol. 5, and hope to have some of the operational manuals available in the coming month. The operations manuals consist of large circuit diagrams, folded into a book-format, and are taking a little longer to scan than the average book!
Other resources:
There are many books, websites, and histories of computing which have more on ENIAC – here are just a few. If you would like to see pieces of the original ENIAC, there are many in institutions across the country including at the School of Engineering and Applied Science at the University of Pennsylvania, National Museum of American History, and the Computer History Museum in Mountain View, California.
I highly recommend this transcript of 50 Years of Army Computing From ENIAC to MSRC (pdf, opens in new window) edited by Thomas J. Bergin from a 1996 symposium. It has fascinating first-person accounts of what it was like to work on ENIAC.
University of Pennsylvania Library has an older, but quite informative history of John Mauchly and the development of ENIAC.
And of course, Landmarks in Digital Computing: A Smithsonian Pictorial History, by Peggy A. Kidwell and Paul E. Ceruzzi, Smithsonian Institution Press, 1994 is pretty good too…
—Keri Thompson
2 Comments
Keri:
I appreciate you covering this topic, but there are some inaccuracies in this account that deserve correction. My mother, Jean Jennings Bartik, just passed away on March 23rd, and I am reviewing her memoirs, which deal with many of these issues.
1. First, I think it is somewhat odd to refer to the women as “operators” and to refer to them as programming by simply plugging in cables and turning function switches.
This is incorrect. The main task of programming then, as it is now, is to figure out the logical sequence of operations needed to be done to solve the problem at hand. Once this was done, because the ENIAC was not a stored program computer, the logical flow of the programming instructions had to be implemented by for example making sure the output of one unit went into the input of another unit, etc. But treating the task of programming the ENIAC as fundamentally one of physical operations of the cables is not at all accurate. Actually, the logic of figuring out how to program the ENIAC was further complicated because the original ENIAC was a parallel processing machine, not a serial processing machine, and the time to solve some of the processes could not be predicted in advance, so the programming logic had to include various checks to synchronize the various diverse calculations going on simultaneously in different portions of the computer.
In calling these programmers “operators”, and referring to programming in terms of its physical manifestations, I think you are reproducing what some of the male engineers associated with the ENIAC might have thought when it was first designed, and why there was willingness to have women do what was thought to be a more minor and more tedious task. Yet programming ended up involving far more intellectual content than that from the very beginning. For example, the programmers of the ENIAC helped the engineers debug the machine down almost to the level of the individual vacuum tubes.
2. Adele Goldsine was NOT in charge of converting the ENIAC to a stored program computer. My mother was. Or rather, she was in charge of a team that did so. Also heavily involved in this effort was Dick Clippinger. This team did go every few weeks to Princeton to meet with Johnny von Neumann, Adele Goldstine, and Herman Goldstine to get advice on programming the instruction set.
Regards,
Tim Bartik
Mr. Bartik,
Please accept first my condolences on your Mother’s passing, and second, my thanks for your corrections and additions to my post on the ENIAC.
In my haste to write a brief post, I condensed and simplified some points that should have been elaborated on, got some facts totally wrong, and obviously stated things in a way that failed to communicate the correct facts I did have.
I have used your information, and have referred back to my original source material, to correct my original post.
Please know that I did not intend to imply by the use of the term ‘operator’ that the women who programmed ENIAC were not performing intellectually challenging work. Using the term ‘operator’ does greatly over-simplify what they did, and I have changed my original post to reflect their actual job – programmer – and only use operator in quotes, as I found it in sources. The original reference to the physical aspect of programming using cables and switches lacked adequate context to understand how complicated the entire programming process was, and I have elaborated on that process to hopefully give a more accurate impression of what programming entailed. I mentioned it only to emphasize that programming ENIAC was not the same kind of programming that most are familiar with today, using high-level languages several times removed from the basic logical functions that the computer performs. I must confess, the descriptions of the complicated process that they went through to program and use ENIAC was a bit over my head! I am completely amazed at what the women who worked on ENIAC did – particularly when I read of them having to figure out how it worked just by looking at the circuit diagrams.
I apologize particularly about the mis-attribution of the stored program process to Adele Goldstine rather than Jean Bartik. That was bad note taking on my part, and I have corrected that reference, replacing it with a list of people involved in the stored program process found on p.47 of the “50 Years of Army Computing….” document I link to at the end of my post.
Sincerely,
Keri Thompson