Go to Antique Computer home page

IBM vs Seven Dwarfs
Why did IBM win?

a log of e-mails and articles

Goal of this page Provide folks with some un-published opinions, hopefully many from those who were present -
This web page started July 23, 2011
Comments, stories, opinions, ... solicited. Send e-mail to Ed Thelen

Table of Contents


Introduction - by Ed Thelen
In the beginning, shortly after Adam and Eve ;-)), people wanted to count things such as sheep, family, days, ...
If you started today herding 13 sheep, you hoped that you would have 13 (or more) tomorrow, and not lose one or more to the neighbors or wolves.
So numbers for counting were invented,
- then addition, six sheep and 13 sheep are ??
- then the bankers and tax collectors wanted to subtract ;-(
- later came multiplication and division.
To ease these operations the Hindus or Chinese invented decimal positional notation,
- the abacus followed.
But numbers were used for more and more things, such as navigation and engineering and astronomy - so mechanical computers were perfected.
- such as logarithms, slide rules, Babbage engines, desk calculators, ...

Census and business requirements involved large numbers of transactions which needed a "paper trail". From about 1890 to 1950 Herman Hollerith's organization, merging into IBM, used punched cards and "unit record equipment" which satisfied that field.

From the mid 1930s to 1950, various inventors, Zuse, Atanasoff, Eckert & Mauchly, Kilburn & Williams, ... made increasingly useful, one of a kind, scientific computing machines. (Little input, lots of arithmetic, and a little output.)

In 1951, Eckert & Mauchly, re-funded/re-organized as the "Univac Division of Remington Rand" delivered the first commercial stored program computer to the U.S. Census Bureau, with plans to market and manufacture more.
Please note the qualifier "commercial", which to me excludes the several ERA 1101 machines developed under contract to the U.S. Navy Bea. - Randy Weaver wonders if this is commercial? - a little about the ATLAS -
This same year, the Ferranti Mark 1 was delivered to Manchester University - but a change in government canceled further British government procurements of expensive computers. :-(( Leaving the second Ferranti machine with no customer :-((

In any case, by 1954, Univac (with eight installations) was the clear leader in electronic computer manufacturer and sales in the United States.

HOWEVER, By 1960, IBM was the clear leader in electronic computer manufacturing and sales in the United States and the world!!.
- Univac and every other computer company, and there were many, were so far behind in sales that the largest were called "the seven dwarfs".
- What happened in those six years to bring about this sudden major change?

Many people have wondered and expressed ideas about what caused this revolutionary change in just six years. I worked for two of the seven dwarfs, and am curious also.


How Did IBM Get Into the Computer Business So Quickly?
from Jim Strickland, July 24, 2011

How Did IBM Get Into the Computer Business So Quickly?

Jim Strickland - Jlstrick @ aol . com

When Remington Rand bought Eckert and Mauchly Computer Company in 1950 they had a two to five year lead over any competitor and some extremely talented personnel. It looked as though IBM was stuck in the mechanical era with its support of the Harvard Mark I relay computer and then its successor, the SSEC (Selective Sequence Electronic Calculator). Further, when Remington Rand bought Electronic Research Associates (ERA) in 1952 it seemed to add to that lead.

Yet within 3 years IBM had not just bypassed Univac, they had put in place the procedures that would make them as dominant in computers as they had been in punched card equipment. How did staid, old IBM progress so fast?

One part of the story starts back in World War II.

IBM was a major contributor to the war effort and had undergone a huge expansion in support of that effort. But Thomas Watson had set a profit margin of 1% on war related work, so money was tight. Watson cut back on research which was not wide ranging to begin with, and he refused to do any research on electronic projects.

But a passionate young engineer named Halsey Dickinson continued to work on vacuum tube electronics in his basement at home. After Watson had a dust-up with Howard Aiken over the announcement of Harvard's Mark I, Dickinson was assigned to prototype IBM's first electronic calculator.

Then, after Tom Watson Jr. visited ENIAC and recommended to his father that IBM not do anything further regarding ENIAC, Tom Jr. visited IBM's Endicott lab in mid-1945. There, Dickinson had linked a high-speed punch card machine to a black metal box about four feet tall. Tom Jr. asked what the box was doing. One of the engineers told him it was multiplying using radio tubes. The engineers explained that it could multiply 10 times faster than the punch card machine could read. In fact, the box spent nine-tenths of its time waiting for the punch card mechanism to catch up.

For some reason, Tom Jr. had failed to see any good in the gigantic ENIAC but Dickinson's electronic black box excited him. "That impressed me as though somebody had hit me on the head with a hammer," he recalled. That moment marked his awakening to electronics and its potential role in IBM products. In Tom Jr.'s memory, he immediately told his father that IBM should put Dickinson's device on the market, and "that is how IBM got into electronics."

The product was the IBM 603 Electronic Multiplier. Announced in September of 1946, it sold over 100 units in short order and confirmed to the Watsons that IBM customers were ready for electronic devices. The 603 was very limited; it could only multiply. But seeing its success, IBM built and announced the 604 which could add, subtract, multiply, divide and crossfoot.

The 604 forced IBM to face, for the first time, a number of problems whose solutions would be tremendously important to IBM and to the computer industry as a whole: the problems of manufacturing, testing, maintaining, and repairing electronic machines which would be produced by the thousands and installed wherever punched-card machines were used.

To manufacture the anticipated volume of 604's, IBM set up the first electronic “clean room”, established an industry first assembly line and developed the first pluggable unit. Those manufacturing (and servicing) innovations not only allowed IBM to sell over 5,600 604's in 10 years, but it set the stage for IBM to be able to mass produce its first computers instead of hand building each one.

So when Tom Jr., early in 1952, went on the tour that resulted in 18 orders for the half-million dollar Defense Calculator (later delivered as the 701), the assembly line process was in place. IBM sold nineteen 701s in three years and over 500 of the lower cost 650 in just five years.) Based on the manufacturing and servicing procedures initiated for the 604, IBM became the dominant computer company, surpassing Univac by the end of 1954.

Success has many fathers, but IBM's fast and successful entry into the computer market was due in no small part to its ability to build on, and adapt, its manufacturing processes to the new electronic era.

Randy Weaver suggests Charles River IBM - Trial Documents
from www.bitsavers.org
US_vs_IBM_Exhibit_14971_part_1_Jul80 - pages 1 - 268, Local copy 14 MB .pdf

US_vs_IBM_Exhibit_14971_part_2_Jul80 - pages 269 - 952 - Local copy 42 MB .pdf

US_vs_IBM_Exhibit_14971_part_3_Jul80 - pages 953 - 1527 - Local copy 30 MB .pdf




A Post Script, a tale from 1975 - by Ed Thelen
By 1975 I had worked for two of the "Seven Dwarfs", General Electric Computer Department and Control Data. Quiting both in turn as they each didn't seem to know how to make money, and I didn't want to be stranded. I was working for Measurex, booming company in paper making process control. I heard that Measurex was going to get a data processing machine, either from "some company" for a "good" price or an IBM System 3 at a much higher price.

I complained to a friend, LaFarr Stuart, who had also worked at a "Dwarf" (RCA), that Measurex might buy IBM. He said that sounded good to him.
I nearly swallowed my teeth !!, paused in shock, and asked why. He said:

  1. IBM knows how to print and handle paper. IBM printers are the most convenient for operators in the industry.
  2. You know IBM will be here tomorrow, and 10 years from now. How many startup computer integrators (computer from one company, printer from another, tape units or disks from a third, software from ???) will survive two years, when you need service?
  3. Payroll is due out tomorrow. Something is sick with the computer. Missing a payroll has a serious impact on all employees. How soon is that computer integrator going to get your machine (from multiple manufacturers) up and running???
I recovered from the shock -
and in fact that IBM competitor was out of business within two years - Measurex would have been stuck with an orphan :-((


If you have comments or suggestions, Send e-mail to Ed Thelen
Return to top