The Community for Technology Leaders
RSS Icon
Issue No.01 - January-March (2000 vol.22)
pp: 62-81
<p>The lines of computing machines that had their origin in the days immediately preceding World War II include a series of calculators Howard Aiken, a professor of applied mathematics at Harvard University, designed. Starting with the Mark I in 1944, Aiken spearheaded an effort that provided not only the physical means of computation but also the tools to direct them and the people to operate them. The third in this sequence of machines was an innovation in design and implementation, while at the same time being conservative in the selection of components. The Harvard Mark III Calculator had the potential to be a significant entry into the field of computing, but events slowed its completion until competitors finished other markedly superior systems. The Mark III was not a machine that would be emulated or replicated beyond its lifetime, but the people who planned it, built it, programmed it, and operated it went on to make significant contributions to the science and practice of computing.</p>
John A.N. Lee, "Howard Aiken's Third Machine: The Harvard Mark III Calculator or Aiken-Dahlgren Electronic Calculator", IEEE Annals of the History of Computing, vol.22, no. 1, pp. 62-81, January-March 2000, doi:10.1109/85.815467
1. Prepared by the 1983 National Computer Conference Pioneer Day Committee.
2. ONR Newsletter, Jan. 1954. [Anon. 1954-1957. Digital Computer Newsletter, Office of Naval Research, Mathematical Sciences Division, Vols. 6-9. (Commonly called the ONR Newsletter.)]
3. ONR Newsletter, Oct. 1954.
4. "Registers," as used in the publications of the time, did not refer to special (usually single-word) storage units within the CPU, but instead referred to elements within the general storage system.
5. The reported diameter of the storage drums (see References 8 and 22) was eight inches, and the circumference was stated as 24 inches. We have been unable to recover the original drums and determine whether the drums were actually eight inches in diameter and the circumference 25.13 inches or 7.64 inches in diameter and 24 inches in circumference. We did recover some of the storage drum pole piece brackets, and there appears to be sufficient room to accommodate an eight-inch drum. It would have been more convenient, however, to have a drum with a circumference of 24 inches, so that there would be room for exactly 12 words (two of which were blank).
6. George R. Stibitz, The Zeroth Generation, unpublished mo nograph, 1993.
7. Computation Laboratory, Description of a Magnetic Drum Calculator.Cambridge, Mass.: Harvard Univ. Press, 1952. (Vol. 25, Annals of the Computation Center.)
8. On the Instructional Tape Preparation Unit.
9. From a group interview of Mark III pioneers by J.A.N. Lee and Henry S. Tropp, held at Mary Washington College, Virginia, 26 January 1986.
10. See N. Stern, From ENIAC to UNIVAC: An Appraisal of the Eckert-Mauchly Computer.Bedford, Mass.: Digital Press, 1981, p. 156.
11. R.L. Wexelblat, History of Programming Languages.New York: Academic Press, 1981, p. 8.
12. An anonymous reviewer pointed out that Konrad Zuse conceived of a similar auxiliary device, the "Plan Preparation machine," that was designed as an attachment to the Z4 computer.
13. M.V. Wilkes, D.J. Wheeler, and S. Gill, The Preparation of Programs for an Electronic Digital Computer.Cambridge, Mass.: Addison-Wesley, 1951.
14. Later Commander; Grace Murray Hopper was always concerned that she reached ranks higher than her mentor did, since she felt that his contributions were more significant (private communication).
15. Hopper, 1981.
16. During the 1986 interviews, Gene Gleissner and Ralph Niemann recalled that the Mark II operator who found the bug and taped it into the logbook was Bill Burke, who later moved to Dahlgren as a computer operator.
17. The K-lab of 1950 is not the same building as is designated as K-lab at the Naval Surface Warfare Center, Dahlgren today. The original building was much closer to the firing range; it was located where it was so that remote sensing signals could be monitored more easily. The designation "K" was chosen for the computer laboratory in 1947 since "C" had already been used for the Command Support Department. Today, K-lab is responsible for Strategic Systems.
18. G.E. Poorte, "The Operation and Logic of the Mark III Electronic Calculator in View of Operating Experience," Rev. Electronic Digital Computers, Joint AIEE-IRE Computer Conf., AIEE, New York, Feb. 1952, pp. 50-56.
19. I.L. Auerbach, "Howard Aiken," Annals of the History of Computing, Vol. 7, No. 4, pp. 354-355, 1985.
20. P. Bird, LEO, the First Business Computer, Hasler Publishing, Wokingham, 1994.
21. Poorte, 1952.
22. Henry S. Tropp and J.A.N. Lee interviews of NSWC participants, Mary Washington College, Jan.26 1986, unpublished.
23. R.L. Wexelblat, History of Programming Languages.New York: Academic Press, 1981.
24. G.M. Hopper, "The First Bug," Annals of the History of Computing, Vol. 3, No. 3, pp. 283ff, 1981.
25. E. Tomash, and A.A. Cohen, "The Birth of an ERA," Annals of the History of Computing, Vol. 1, No. 2, pp. 83-97, 1979.
26. M.R. Williams, "Howard Aiken and the Harvard Computation Center, Meetings in Retrospect," Annals of the History of Computing, Vol. 6, No. 2, pp. 157ff, 1984.
27. The reprint of this paper states that it was presented in Pittsburgh, Pennsylvania, however other references to the 1952 ACM conference also list Toronto as the venue.
28. Anon., July 1950. A Survey of Large-Scale Digital Computers and Computer Projects, Office of Naval Research, Washington, D.C., declassified Oct. 1985.
29. The ONR report lists the high-speed memory as 100 words, though records at Dahlgren NSWC state 96 words. Since the Mark II was actually two separate machines that could be operated either in tandem or together, it is not clear whether this is the total for both machines or individually.
30. The Mark III had 4,000 words for instructions and 4,350 words for data storage.
31. Also known as the BTL Mark IV.
32. Six machines were in construction, with the first to be delivered to the customer in 1951.
33. Provision was made to extend the memory to 3,000 words later.
34. Later modified to 0.2 millisecond using the CRT electrostatic memory.
35. Later extended to 4,096 words.
36. In construction.
37. Also known as the Manchester Automatic Digital Machine (MADM) or Manchester University Computer (MUC). The commercial version, also in construction in 1950, was the Ferranti Mark I.
12 ms
(Ver 2.0)

Marketing Automation Platform Marketing Automation Tool