Software Development at the Eckert-Mauchly Computer Company Between 1947 and 1955
Yüklə 1,23 Mb. Pdf görüntüsü
|
Iterations – Norberg – Software Development at EMCC 20
applications personnel. In September 1955, Mary K. Hawes, Supervisor of Commercial Programming of Remington Rand, presented a talk entitled “Automatic Routines for Commercial Installations” at a meeting of the ACM. After 1955, attendance at meetings by Remington Rand employees became too numerous to catalog here for any useful purpose.
Figure 4. Pseudocode instructions for A-0.
Iterations – Norberg – Software Development at EMCC 21
Figure 5. Flow-matic Code.
All of these examples of codes, compilers, and outreach indicate the high level of software activity within EMCC and later Remington Rand. The company assembled a group of highly effective programmers to provide programs that would make the UNIVAC more attractive to potential customers and to attach customers to Remington Rand. The group was effective, and after the sales force received education about the use of computers, sales began to rise, such that the late 1950s and 1960s ensured the future of Sperry Rand in the computer field.
Almost immediately upon EMCC’s joining Remington Rand, difficulties about programming arose within the firm. Remington Rand inaugurated a Department of Program Planning (outside of the Eckert-Mauchly subsidiary) connected with the sales activity, and hired personnel from Harvard Computation Laboratory, Aberdeen Proving Ground, Institute for Advanced Study, and Dahlgren Proving Ground, all people with experience with computers. In addition, meetings about programming training with senior management of Remington Rand by EMCC representatives began on May 5, 1950, when Mitchell, Wistar Brown and Mauchly met with Al Seares to discuss the need for a new training program. Noting that this activity had some urgency to it, EMCC offered cooperation. Brown repeated this plea in a second meeting on May 31
Iterations – Norberg – Software Development at EMCC 22
with Millang. Nothing happened for the next two months. In August, members of Department of Program Planning (DPP) visited Philadelphia for six weeks of training, without consultation or indication of purpose, after which they were recalled to New York. On September 1 st , David
Savidge was appointed to head Program Planning. No mention was made of EMCC. In the fall of 1950, Herbert Mitchell learned that this department had undertaken to “invent, develop and promote a different method of programming” and had organized a course in which they advertised this new method for programming the UNIVAC. Betty Snyder visited the Census Bureau on November 2 nd and found advertising materials from DPP sent to Census by Savidge. Mauchly asserted that the advertising materials circulated were misleading and contained serious omissions and errors. He telephoned Millang requesting an appointment to discuss the materials. Millang promised to call back, but he never did. In November, the Philadelphia group also learned from a customer that a “trial course” date had been set. This intelligence set in motion a chain of telephone calls and visits to Philadelphia. Savidge came and met with Mitchell and Hopper. Mitchell and Hopper noted that every page of the text material Savidge brought with him needed correction. When asked about the purpose of the course and the new methods, Savidge invited them to attend the course, though without answering their question. Mitchell (and Mauchly for two sessions) visited the course. The students in the course believed the methods were those of EMCC. Mauchly called for a test of the methods developed by DPP and the EMCC Computation Analysis Laboratory, successor to the Applications Group. This entire contretemps was remarkable when one considers that the DPP group had access only to descriptions of the UNIVAC. Mauchly carefully criticized the materials and DPP, noting that the DPP group was composed of knowledgeable and experienced people, and all he wanted to do was cooperate with them to improve the process, the materials, and training of people to use UNIVAC, and promote Remington Rand. 54 The records do not show how this problem was resolved, but the problem is symptomatic of Remington Rand management’s approach to its new subsidiary. 55
After the acceptance of UNIVAC I by the Census Bureau in March 1951, the Eckert-Mauchly Division began to issue UNIVAC system information in the Remington Rand style. The company published booklets describing the UNIVAC system, problems that could be solved using it and the software developed by EMCC, and the availability of training programs designed to teach customer personnel how to operate and program the system. Besides adopting names for compilers like Math-Matic and Flow- Matic, EMCC described their product as the “UNIVAC Fac-tronic System.” This description presented information on the capability of the parts of the system and how they related to each other, the reliability of the system, sorting with UNIVAC, the range of applications programs availability, and the seminars and training activities designed for the Iterations – Norberg – Software Development at EMCC 23
customer. 56 The specific applications discussed in detail were used at the Census, materials control in manufacturing, from the development of a production schedule to service schedules, and payroll preparation. EMCC indicated that a wider range of applications was possible with the Fac- tronic system. In the commercial area, they cited programs for billing, sorting, collating, interfiling, nearly 100 statistical and accounting report possibilities, and all the elements of reporting taxes, social security, and deductions necessary in payroll accounts. EMCC had developed programs for statistical analysis for both military and civilian users. Various logistics programs included production scheduling, building requirements, stock control, etc. The UNIVAC could also be used for scientific and engineering applications, such as the solution of matrix algebra problems, several elliptic partial differential equations, including LaPlace’s and Poisson’s equations with various shaped boundaries and boundary values, and the rapid reduction of test data from experiments. The standard library programs could be compiled using Math-Matic and Flow-Matic in the mid-1950s. The customer’s personnel could design specialized programs using some of the library of routines developed by EMCC, after they mastered the system in the training courses offered by Remington Rand.
All of these programs, both systems software and applications, had been developed by the mid-1950s by the various programming groups of EMCC and Remington Rand to be delivered with the UNIVAC I. Sometimes the customer could not wait for Remington Rand to produce programs they needed for their operations. When this happened, the customer using systems like Flow-Matic could develop the needed programs. EMCC and Remington Rand were very conscious that sales of UNIVAC would only happen if applications programs came with the UNIVAC system, and the company expended a substantial effort in the years 1947 to 1956 on coding techniques and programming systems.
Arthur L. Norberg, “Software Development at the Eckert-Mauchly Computer Company Between 1947 and 1955,” Iterations: An Interdisciplinary Journal of Software History 2 (December 31, 2003): 1-26.
1 See, for example, Paul E. Ceruzzi, A History of Modern Computing, (Cambridge, MA: MIT Press, 1998), Chapter 3; Martin Campbell-Kelly, “Programming the EDSAC: Early Programming Activity at the University of Cambridge,” Annals of the History of Computing, 2(1980): 7-36 and
Shapiro, “Computer Software as Technology: An Examination of Technological Development,” Ph.D. dissertation, 1990, Carnegie-Mellon University, Chapter 2. 2 EMCC employed the name EDVAC II to distinguish their company design for a computer system from the University of Pennsylvania Moore Iterations – Norberg – Software Development at EMCC 24
School EDVAC design, which they helped design before leaving the Moore School in 1946. Later, the EDVAC II became the UNIVAC I. 3 W. Barkley Fritz, “The Women of ENIAC,” Annals of the History of Computing, 18(Fall 1996): 13-28. 4 UNIVAC Conference Transcript, OH 200, CBI. Comment by Frances E. Holberton, p. 52. 5 Fritz, “The Women of ENIAC,” op. cit., p. 18-19. This article contains substantial information about Bartik, her training, and the programming of ENIAC.
6 Interview with Frances E. Holberton, OH 50, passim, CBI. 7 UNIVAC Conference, op. cit., p. 65. 8 Ibid., p. 68. 9 T. W. Brown to H. L. Strauss, “Report on Applications Dept.,” 31 March 1949, Sperry Corporation Records, Acquisition 1825, Box 83, Chronological File, Hagley. 10 Association f or Computing Machinery, “A Quarter-Century View, ACM71,” (New York, 1971), and “An Analysis of the Eckert-Mauchly Computer Corporation,” op. cit., Biographical Summary Section. 11 Ibid.
12 “Conferences on EDVAC II Design,” March 11, 12, 1947, Sperry Corporation Records, Acquisition 1825, Box 37, Hagley. 13 Ibid. 14 Ibid. 15 Ibid.
16 Data taken from a comparative chart in the Frances E. Holberton Papers, CBI 94, Box 13, File: UNIVAC Code Development. 17 Ibid. 18 Ibid.
19 Nancy Stern, From ENIAC to UNIVAC: An Appraisal of the Eckert- Mauchly Computers, (Bedford, MA: Digital Press, 1981), p. 133. 20 C. J. Bashe, L. R. Johnson, J. H. Palmer, and E. W. Pugh, IBM’s EarlyComputers, (Cambridge, MA: MIT Press, 1986), pp. 138-142. 21 Ibid., pp. 143-44. 22 The Holberton papers at CBI contain outlines and some lectures from the course offered to EMCC engineers in early 1950, with lectures by Herbert Mitchell, Grace Hopper, and Betty Snyder. 23 “Training Course for EMCC’s Engineers,” spring 1950, Holberton Papers, Box 5. 24 Mauchly, “Chronology,” October/November 1947, Mauchly Papers, University of Pennsylvania Archives, Box 3:C:1, Folder 5. 25 Lukoff, From Dits to Bits, op. cit., passim on Eckert’s presence at the office and how he interacted with personnel. 26 B. Snyder and H. M. Livingston, “Coding of a Laplace Boundary Value Program for the UNIVAC,” MTAC, 3(January 1949): 341-50. Iterations – Norberg – Software Development at EMCC 25
27 H. Liebmann, “Die ausgenährte Ermittlung harmonischer Funktionen und konformer Abbildungen (nach Ideen von Boltzmann and Jacobi),” Adad. D. Wissen., Munich, Berichte, 1918, pp. 385-416. 28 “Coding of a Laplace Boundary Value Problem,” EMCC 1948, Holberton Papers, Box 13. 29 G. M. Hopper, “The Education of a Computer,” Proceedings of the Association for Computing Machinery (Pittsburgh meeting, May 2 and 3, 1952), (New York: ACM, 1952), pp. 243-249. Hopper used this title for several presentations in this period, but she claimed the presentations differed. A comparison of her published paper “The Education of a Computer,” with “The Education of a Computer” in Proceedings, Symposium on Industrial Applications of Automatic Computing Equipment, Midwest Research Institute, Kansas City, Missouri, January 8 and 9, 1953 (copy at CBI In Hopper file), illustrates that the basis of the many papers was the description of the UNIVAC system, but the examples of what the system could calculate were chosen to appeal to the audience Hopper was addressing. The description in the text can be found in either article. 30 D. Knuth and L. T. Prado, “The Early Development of Programming Languages,” Encyclopedia of Computer Science and Technology, 7 (1977): 419-493. 31 “Early American ‘Compilers’,” Ibid. 32 M. V. Wilkes, D. J. Wheeler, and S. Gill, The Preparation of Programs for an Electronic Digital Computer, with special reference to the EDSAC and the use of a library of subroutines, (Cambridge, MA: Addison- Wesley, 1951; reprint edition, Los Angeles: Tomash Publishers, 1982). See Martin Campbell-Kelly’s introduction to the reprint edition for an assessment of the significance of this publication. 33 Knuth and Pardo, op. cit., p 434. 34 “UNIVAC Short Code [Instruction Manual],” “Preface,” Computer Product Manuals Collection, CBI 60, Box 186. 35 “UNIVAC Short Code,” passim, and Sammet, Programming Languages, op. cit., pp. 129-130. 36 G. M. Hopper, “Keynote Address,” In Richard L. Wexelblat, ed., History of Programming Languages, (New York: Academic Press, 1981), pp. 7-20, p. 10. 37 Jean E. Sammet, Programming Languages: History and Fundamentals, (Englewood Cliffs, NJ: Prentice-Hall, 1969), p. 12. 38 Ibid. 39 The following description is taken from the Instruction manual “A-0 Compiler,” Computer Products Manual Collection, CBI 60, Box 205. 40 Hopper, “Keynote,” op. cit., p. 11. 41 “A-0 Compiler” Instruction Manual, op. cit., p, 3. 42 Hopper, “Keynote,” p. 12. Iterations – Norberg – Software Development at EMCC 26
43 Richard K. Ridgway, “Compiling Routines,” presented at ACM meeting September 8-9, 1952, Holberton Papers, CBI 94, Box 5. 44 L. Stowe, “Programming,” Summary of Papers Presented at the Seminar on Data Handling and Automatic Computer, 26 February to 6 March 1951, Office of Naval Research, US Government Computing Collection, CBI 63, Box 2. 45 “Progress Report on Bureau of the Census Problem, 4 January 1951,” Holberton Papers, CBI 94, Box 23. Stowe, “Programming,” p. 80. 46 Ibid., p. 81. 47 Ibid.
48 Ibid., p. 83. 49 Hopper’s group included James McGarvey, Adele “Millie” Koss, F. M. Delaney, Margaret H. Harper, and Richard K. Ridgway. Hopper, “Keynote,” op. cit., pp. 12-13. 50 The following description comes from a Remington Rand “Instruction Manual: The A-2 Compiler System” published in 1955 In the Holberton Papers, CBI 94, Box 18. 51 Hopper, “Keynote,” op. cit., p. 14. 52 Proceedings, Harvard 1948; reprinted as Volume 7 In the CBI Reprint Series, MIT Press and Tomash Publishers, 1985. 53 Mauchly Papers, Box 3:C:12, Folder 258, University of Pennsylvania Archives. 54 Mauchly to A. N. Seares, Internal evidence points to a series of memoranda written toward the very end of 1950, Mauchly Papers, Box 3:C:6, Folder 140, University of Pennsylvania Archives. 55 When Remington Rand agreed to purchase EMCC, it was with the understanding that EMCC would be an independent part of the Remington Rand company, i.e., a subsidiary, reporting directly to the head of the company James H. Rand. 56 One of many examples that can be found in the Computer Product Literature Collection (CBI 12) is a 20-page brochure “UNIVAC Fac- tronic System by Remington Rand Inc. Eckert-Mauchly Division,” Box 99. While there is no date on the brochure, internal evidence about available subsystems suggests 1953. Yüklə 1,23 Mb. Dostları ilə paylaş:
|