Home Page
The following article was taken from the
“Annals of the History of Computing”, Vol.2  #3  July 1980

The author, Fred Gruenberger, worked at the Rand Corporation for eight years. Until his retirement, he was a professor of computer science at California State University, Northridge, and editor of the magazine Popular Computing.

This is a written report of a talk January 17, 1958, at the Digital Computer Association, a local informal group. The Southern California area is  today generally regarded as a center of activity in high speed computing, having perhaps the highest  density of machines and active, prominent people of any area in the world; the rise of this intense activity is traced from 1942 to 1957.

The Southern California area is today generally regarded as a center of activity in high speed computing, having perhaps the highest density of machines and active, prominent people of any area in the world. The development of this intense activity is  worth recording, especially in view of the fact that much of the pioneering in the field was done elsewhere.

A few words of definition. The area in question centers around Los Angeles and extends north to the Naval Ordnance Test Station at China Lake and  south to San Diego. A third vertex of the triangular area is marked by Point Mugu where the RAYDAC is located.

Second, it becomes necessary to define computing.  The distinction between what was or was not considered computing in the early days rested on two criteria: was it automatic (that is, could the operator walk away from the machine and have the process continue?), and did it involve multiplication and/or division? Thus, application of the Mendenhall-Warren-Hollerith method of progressive digitizing for calculating sums of products on a tabulator would count as a computing application; large, complicated punched-card tabulating systems would not. The discussion here concerns only digital equipment.

In preparing this account, there were two sources of information. First, people: people checked in from all sides with anecdotes, claims of having the first of a given machine, leads to other people, and an urgent desire to get the record straight. The second source was the written records of the time. Several installations in the area came forth with volumes of memos dating back ten years and more.

A good source of factual information was the collected proceedings of the various IBM Research forums. The first of these meetings, in 1940, listed no attendees or speakers from California; it was obviously a meeting of Eastern blue-sky dreamers. The next one, in 1946, followed the same pattern. The third forum, in 1947, attracted two persons from this area, bold souls who crossed the plains to see what was cooking. (This was an Educational Research Forum, attended by Georgia Adams of the Pasadena Public Schools and C. Lorene Fitch of the Glendale City Schools.) The 1948 forum drew a crowd of seven from California, four of whom presented papers. Bill Bell, of  Telecomputing, described an application on the 602. Gertrude Blanch, of INA, discussed differencing on the 405. Ben Ferber, of Convair, presented a paper entitled "Planning Engineering Calculations for IBM Equipment." And George Fenn, of Northrop, described a Rube Goldberg marriage of the 603 and 405 machines, which combination was said to form a powerful computing tool.  Recall the context of the time: a few unique stored-program computers were then in operation, but few people were dreaming of any organized mass-production attack on come, man problems.

The 1948 forum was the first with the label "scientific computation." In addition to the speakers just mentioned, the meetings were attended by Paul Bisch, of North American Aviation, C. H. Stevenson, of Douglas Aircraft, and Harley Tillitt, of NOTS, China Lake.

The November 1949 forum was attended by Bruce Oldfield, of NOTS, China Lake, and by Verner Schomaker, of the California Institute of Technology. Herman Kahn, of The RAND Corporation, spoke on "Modification of the Monte Carlo Method." William Woodbury, of Northrop, described uses of the 603-405 "Combo." At this forum, also, came the announcement of the production version of this gadget; the CPC-I.

At the December 1949 forum, things began to pick up. The Southern California aircraft industry had obviously begun to discover that computing machines were useful; six speakers were from this area, and four other people attended. The six who spoke (Everett Yowell, of the Institute for Numerical Analysis: Paul Bisch. of North American Aviation; John Lowe, of Douglas Aircraft; Cecil Hastings, Jr., of The RAND Corporation; Bill Bell, of the Telecomputing Corporation; and Gregory Toben, of Northrop Aircraft, Incorporated) accounted for eight papers.  Both Yowell and Lowe gave two papers. Uses for the 602, 602A, 604, and the CPC were described. The significant advance was made by Bill Bell, who spoke about the DUZ panel for the 604—the first practical approach to general-purpose computing, using the available mass-produced equipment. It was at the forum that Cecil Hastings first presented his work on approximations, which developed into a significant contribution to the computing field.

At the 1951 forum, there were six talks from Southern California people, all on uses of the CPC One might say that this area had now joined the club and was starting to assume a position of leadership. UNIVAC I had been out for two years by this time. The six papers were presented by Murray Lesser, of Northrop Aircraft, John "Don" Madden of  The RAND Corporation, John Lowe, of Douglas Aircraft, Bill Bell, of Telecomputing, Everett Yowell of INA, and (combined) Harley Tillitt, Martha Kenyon, and Bruce Oldfield, all of NOTS, China Lake.

It should be noted, if one detects a strong IBM flavor to this account, that Southern California has not been particularly good Remington Rand territory. The first Remington Rand computer in this area was a 409, installed in June 1952 at Douglas Aircraft. The first large Remington Rand machines were 1103s, one at Convair and one at Ramo-Wooldridge, in late 1954. The first UNIVAC I was installed at Pacific Mutual Life Insurance in August 1955, and a UNIVAC Scientific (1103A) was installed at Ramo-Wooldridge in April 1957.

But to return to the more ancient history. Prior to the fall of 1945, some computing was being done, using punched-card equipment. Significant work was lone by Ernest C. Bower, as a Douglas employee, at the Cal Tech Cooperative Wind Tunnel, and by Bill Bell and John Lowe at Lockheed.  Stanley Rogers had performed some electrical load analyses at Convair as early as 1942.

In the fall of 1945, the 601s started being installed in the West Coast area (they were in use in the East some time prior to this). This machine, you may recall, was a relay device that could read two factors punched in one card and punch their product into the same card. The early 601s had no control over minus or plus signs and, being rather slow, answered  superbly the requirement of allowing the operator to eve the machine for long periods.

All the people who were active at that time agree that a significant advance in the available hardware occurred in May 1946. At that time the feature of sign control was added to the 601. Nothing but stored - programming seems to have equaled the impact of that improvement, at least in the minds of those who were there.  Prior to the production version, Bill Bell and Tom Lowther, at Lockheed, had had a 601 modified to include sign control.

Late in 1946 a digital computer group was formed at Northrop. The name of Northrop pops up again and again in any computing history; this is one of the early references to their foresight.  In 1946, few people bothered to distinguish between digital and analog, and in most places computing was either a side job of the tabulating department, or was being done clandestinely (clandestine: conducted with secrecy, usually for an evil purpose; illicitly covert) by some renegade engineers. Northrop was facing up to the future.

In December 1946, the Project RAND group of Douglas Aircraft was wondering how to go about punching the output from a device they had constructed to generate random digits. Cecil Hastings, Jr., who was head of the computing department at that time, noted in a memo that two possible choices were:
· an 016 key punch; rental $18.00/month
· a 516 summary punch; rental $30.00/month
His memo says, "I felt that the difference in rental would almost certainly rule the type 516 out."

In 1947, Northrop contracted with the Eckert-Mauchly Corporation to produce the BINAC. In May 1947, a meeting was called of those persons interested in engineering computations to discuss common problems. The roll call of that meeting is a star-studded "who's who" of the pioneers, many of whom are still very active in the field. It is appended to this history as an artifact.

Later in 1947 the 602s started arriving. Since this machine could be programmed, it rates as one of the first mass-produced automatic calculators. A RAND memo, dated only three weeks after the arrival of their first two 602s, stated, "We've got to have two more of these machines!" The quote epitomizes the whole business.

The year 1948 was marked mainly by the announcement of the 602A and 604 by IBM; by mid-1949 both machines were arriving in droves in this area. A memo of the time, commenting on the profusion of equipment and the new ideas, noted, "in another year it will be impossible to break a new person into this department." That memo would not sound strange today.

Another comment of mid-1949 is worth quoting. Word had arrived of the construction at Los Alamos of the MANIAC, with 1024 words of high-speed storage. "To us," the memo read, "this is virtually an infinite memory." The run-of-the-mill installation was, at that time, dreaming of 80-word memories of the about-to-be-delivered CPC-I.

October 1949 was highlighted by the appearance of the first applied science representative of IBM, in the person of Lloyd Hubbard (now a district applied science manager).

It was in 1949 that RAND, having decided that stored-program computing was becoming increasingly important, started design work on the JOHNNIAC (a Princeton-type machine).

In November 1949, the MADDIDA (the first electronic digital differential analyzer) at Northrop became operational. Northrop's contribution to the CPC via their work on the 603-405 combination, and to the UNIVAC via their work on BINAC, has been mentioned.  The work on the 603-405 "Combo" is generally credited to Greg Toben and Bill Woodbury, both of whom are now with IBM. When MADDIDA appeared to be successful, a group of men at Northrop became impatient at the efforts to promote such equipment and spun off in a new company, Computer Research Corporation. CRC was ultimately absorbed by National Cash Register. Northrop meanwhile went ahead with MADDIDA and later sold the rights to it to Bendix. Glenn Hagen, who was in charge of the MADDIDA project, left Northrop later to help form Logistics Research, Inc., which later produced the ALWAC. And Floyd Steele, also of the MADDIDA group, formed Digital Controls, a company later absorbed by Litton Industries. Thus, the digital computer group at Northrop has had a finger in many computing pies, one way or another.

The CPCs arrived en masse in 1950. Northrop received a prototype sometime earlier for their work in development of it. Eight of the production models were shipped in a batch at approximately the same time. In compiling this history, it was difficult to determine who was indeed first on the air with one. The honor seems to belong to RAND, who beat out the Institute for Numerical Analysis by a matter of days. INA can justly claim, however, to having the first general-purpose fixed-point panels.

In June 1950, R. F. Clippinger was quoted in the American Math Monthly as predicting that within ten years as many as 2000 persons would be needed as computer programmers. About the same time, George Brown, now director of the Western Data Processing Center, also stuck his neck out and predicted that the day would come when such people would have to be paid as much as $425 a month. Before anyone laughs too loud, he should put into writing his own predictions for 1970.

By July of 1950 the first floating-point setups for the CPC were operating. The first such setup, largely due to Don Madden (then with RAND) was a 5 digit affair. The write-up for it conjectured that an 8-digit scheme might be feasible.

The SWAC was dedicated at INA in August 1950; other than the ill-fated BINAC, this was the first stored-program machine on the West Coast.

During 1951, the computing fraternity was discussing the "defense calculator" proposal of IBM (which device emerged as the 701 two years later; actually, the 701 was never ''officially announced"— the first machine was delivered to WHQ in March 1953). In December 1951, Computer Research Corporation shipped the first of their CADACS.

Early in 1952 RAND ordered from IBM a dozen tab carriages—no tabulators, just carriages. (They were wanted as a device to flip sheets of preprinted paper in front of subjects, to simulate radar scopes. This was the first glimmering of ideas that led eventually to the System Development Corporation.) Those who can recall those days can appreciate the fun involved in ordering from IBM part of one of their machines. The RAYDAC passed its acceptance tests (at Waltham) in July 1952.

In November 1952, the organizational meeting for the Digital Computer Association was held. Some explanation is in order. Why, with a national organization like the ACM in operation, should groups like the DCA form? (Note that there are at least five such organizations today: besides the DCA, which was the first, there is MC2 in Chicago (Mid-Continent Computer Association); [(DC2)] in Washington (District of Columbia Digital Computer Association); MADCAP in Minneapolis (Minnesota Association for Digital Computing and Programming); Northwest Computer Association in Seattle.)

Perhaps some of the philosophy behind such organizations can be felt by reading excerpts from a letter of R. Blair Smith, of IBM, who was one of the founders of the DCA:

You suggested that I might be able to offer information about the founding of the DCA. I would have to search through my old expense account records in order to determine the exact date of our organizational affair at the Santa Ynez Inn.

At the time, I had recently assumed my first computing territory and was immersed well over my non mathematical ears in dealings with scientific people for the first time in my life. Then, I naively believed that old saw that I would have to become an expert at psycho ceramics which, as you know, concerns the care and feeding of crackpots. Later, I found this fallacious and have since preferred the companionship of people like the members of the Digital Computer Association.

The reasons for organizing the DCA were rather obvious. A number of companies (and more importantly some of my own customers) had ordered and were busily training programmers. Due to the scarcity of trained personnel, it was clear that the 701 customers needed to share information and know-how.

Also, at times there were less than cordial relation when occasionally one company would hire a programmer trained by another company. While I did not believe that the organization of DCA would stop the movement of personnel, I did think that the congenial atmosphere of
the Santa Ynez Inn might promote friendly relations.

In your letter, you honor me as a founder of DCA.  I am proud of the part I played, but I am sure that you must realize that my customers gave me the idea and provided the nucleus of strong support which made DCA successful. For instance, I remember discussing with Paul Armer, of RAND, the need for better communication between 701 customers. While I cannot remember the details of that discussion, I would venture to say that, based on my knowledge o Paul, it was he who suggested that such a group be organized. He probably also suggested that IBM pick up the tab for the first dinner.

DCA, then, stresses first of all good fellowship among computing people, and sponsors incidentally a talk at each dinner meeting on some phase of the art.

The old Selective Sequence Controlled Calculate at IBM's WHQ was dismantled in January 1953 and the first 701 was installed. One equipped with 20/20 hindsight can clearly see that the era of mass-produced big computers was upon us. Yet in the same month there was printed a statement from a local expert (Rex Rice): "Eight to twenty-eight words of high-speed memory is all that one needs."

In June 1953, CRC put into operation a 600-line-per-minute printer, printing 120 characters wide Not bad for 1953. In July, RAYDAC went on the air at Point Mugu. By the end of the year the 701s wee arriving in this area.

In February 1954, the JOHNNIAC went on the air at RAND, using a Selectron memory of 256 words. (The contract had already been let for a core memory for JOHNNIAC.)  Honesty insists that it be recorded that the machine was computing (of all  things) prime numbers then as one of its first test  routines. In August 1954 the first DATAtron was delivered to the Jet Propulsion Laboratory, and later that year, as was previously mentioned, 1103s went  to Convair and Ramo-Wooldridge.

 Late in 1954 an event occurred that was to have national significance. A group of 701 users in the Southern California area decided, at the suggestion of Frank Wagner and Jack Strong, based on the success of the DCA, that (1) a better coding system  for the 701 was needed, and (2) no single installation could produce it alone. The group formed into an  informal organization called PACT (Project for the  Advancement of Coding Techniques) and proceeded  to produce PACT I, a compiler. PACT-IA for the 704 followed later. A complete account of the PACT  efforts can be found in the October 1956 Journal of the Association for Computing Machinery. The point about  PACT was the cooperative effort; a direct result of this effort was the formation of SHARE, the first of the user groups. From SHARE, we have proceeded  to GUIDE, DUO, USE, and so on.

 In February 1955, JOHNNIAC got its 4096-word core memory from International Telemeter (now Telemeter Magnetics); this was the first large, commercially produced core memory. The presence of TMI (the only independent core manufacturer on the West Coast) is itself part of the computing history of this area.

 In 1956 the 704s arrived in this area (the first one  at RAND), and in September the first
705 arrived at Farmers Mutual Insurance Company.

 A UNIVAC File Computer was delivered in April  1957 at Douglas Aircraft, and in May
the first 32,768 word core for the 704 arrived at RAND.

So much for the chronology of the hardware. At  the DCA meeting in January 1958, a group of "old timers" who are still active in the field were invited  as special guests and were each called upon to recount some tale of adventures in Computerland. The group included (with their affiliations at the time they were discussing):
  Charles Davis, North American
  William Bell, Telecomputing
  Jack Strong, North American
  Cecil Hastings, Jr., RAND
  George Brown, RAND
  Ernest Bower, Douglas
  William Gunning, RAND
  Paul Armer, RAND
  Harley Tillitt, NOTS, China Lake
  Ben Ferber, Convair
  Don Eckdahl, Northrop
  Lee Ohlinger, Northrop
  Walt Schleisser, Douglas
  Harold Sarkissian, Northrop
  Dan Sonheim, RAND
  John Lowe, Douglas
  John "Don" Madden, RAND

As one can well imagine, the tales were the 602A that stacked its cards careful window . . . the CPC iceboxes that were used to store ice cream . . . the woman who was commissioned to watch when the counters of the 602 rolled negative (and did watch)....

Despite the jocular and slightly deprecatory air of the accounts furnished by the old-timers, there is little doubt of the value of their pioneering work. It is probably true for the majority of those in computing today that these men were turning out useful answers at a time when the rest of us weren't aware that automatic computing existed.