Introduction

The origins of computing

The story of computing's development is as fascinating as anything in history. In just more than 50 years, we have gone from some sketchy ideas and concepts to a world in which the number of computing devices is reckoned in the hundreds of millions and growing fast.

* Early computing devices

The concept of a mechanical calculator dates to the Sixteenth Century, and was realized in fits and starts in various ways over the succeeding centuries. By the end of the 1800s, companies were producing devices that were sophisticated and reliable. But, even as businesses and scientists came to rely on these machines, it was obvious that the use of gears and levers would always limit their functionality. Most important, it wasn't practically possible to create mechanical devices that could be programmed. You could add a long list of numbers to get a result, then divide that result, etc. But you couldn't tell a mechanical device to add some numbers, compare the result to some other number, then either divide or multiply depending on the outcome of the comparison. You had to have people make the intermediate decisions, which meant that the operations were invariably slow and error-prone.

The idea of an electronic computer surfaced not long after the appearance of electronics. It seemed clear to creative people (this is in retrospect, of course) that the vacuum tube, or more accurately collections of vacuum tubes, could do what mechanical devices couldn't—temporarily store the results of calculations and instructions about new ones. The actual realization of an electronic computer occurred in an American university, though which one is the subject of intense debate.

* The University of Pennsylvania vs. Iowa State

The University of Pennsylvania vs. Iowa State—not a football game, but a controversy over where the first electronic computer was developed. While most sources credit the first electronic computer, ENIAC (1946), to two University of Pennsylvania researchers, John Mauchly and J. Presper Eckert, there is strong evidence that the first machine was actually built in 1942 by a professor at Iowa State, John V. Atanasoff.

Whatever the verdict, we know that the American university has been a critical part of the initial and continuing development of computing. The first computer was built in an American university, and the first computer company was a direct university spinoff. Universities soon created an entirely new discipline to support the fast growing industry, and their classrooms and labs supplied the educated people as well as much of the actual knowledge that has driven an extraordinary pace of change. The university role, however, has always been in partnership with business and government.

* The partnership of IBM and the Department of Defense

IBM was already a mature and well-established company when the first electronic computer was created. Under the leadership of the indomitable Thomas J. Watson, Sr., IBM had taken a position of leadership in office machinery, including devices that performed calculations. The company was caught flat-footed by the advent of the electronic computer, but Watson quickly divined its importance and made an all out effort to secure leadership. There were several ingredients to IBM's rapid success, but none is more important than its understanding of the kind of businesses that would use computers. Even though the technology was new and fragile, IBM appreciated that potential customers, the kind that could afford computers, were not interested in new technology or experimentation. In fact, they were generally risk-averse. A typical IBM customer was a utility company that needed some way to deal with the huge task of calculating, printing, and reconciling its customers' bills. To companies like these, the computer presented an enormous opportunity. As the nation's population grew, and as society experienced a new level of prosperity, the challenge of hiring and housing a vast army of clerks was increasingly difficult—perhaps at some point impossible. On the other hand, the computer was a danger. If just one billing cycle was screwed up, the company would face a disaster of enormous proportions. The company would survive—people would still need electricity—but the executives surely wouldn't. IBM understood this environment perfectly, and provided systems that were amazingly secure and stable given the precariousness of the technology. This ability allowed IBM to dominate its many competitors.

Another dimension of IBM's relationship with its large and conservative customers was that there was no need for dramatic improvements in the technology. If IBM could regularly provide more for less, which was easy to do with a technology that was in its earliest stages, its customers were satisfied. Competitors could and did offer more, but their chronic inability to provide IBM's rock solid reliability and service kept them at the margins. Progress in computing might have continued at a glacial pace were it not for the Department of Defense, which was far more interested in seeing rapid advances incorporated into weaponry and related systems. The Pentagon liked IBM as a partner for the same reason as did the large corporations, and IBM obliged in advancing technology by building, in partnership with US universities, a very strong research capability. As a result, whenever competition forced IBM to pick up the pace on the commercial side, it was ready with something from the lab. This cozy relationship continued until it was broken by the accelerating pace of technological development. To understand why this happened, we need to review the development of computer "generations."

Five generations of computing

Counting computer generations is necessarily controversial—machines don't have the same pedigrees as people. But, the five generations that are described here comprise close to a consensus. We'll characterize them briefly, then discuss the fundamental changes in economics that have resulted.

* The mainframe

The structure of the mainframe hasn't changed a lot since the earliest computers. Its primary characteristic is that all intelligence (computing power), as well as all storage of data and programs, is at the center; kept in the cabinet or cabinets that are the main frame(s). Users get access to the mainframe's intelligence and resources through terminals—dumb devices that are little more than a keyboard and a display.

* The minicomputer

This is probably the most controversial of the generations since it is just a variation on the mainframe—the same basic centralized organization just on a smaller scale and with lower production costs. The reason that the mini is described as a generation is that its lower prices sharply increased access to computing beyond the large corporations that could afford mainframes.

* The microcomputer

The microcomputer, generally synonymous with the personal computer, really is a generation since it offers a dramatic contrast to its predecessors. Where in the past individual users all shared the resources of a single machine, now a single user had direct and personal access to significant computing power as well as to stored programs and data. Eventually, this contrast was blurred as single-user machines were connected with each other over local area networks (LANs), which were then often connected back to mainframes and minis. But, even when all computer users were linked to the big systems, the relationship was fundamentally different. Now, the user had a great deal of independence, and could share with others o

E. GARRISON WALTERS is Vice Chancellor for Academic Affairs with the Ohio Board of Regents. He led initial development of OhioLINK, a statewide computer system for academic library users, and helped plan the expansion of Ohio's Internet-connected statewide computer network to link all public and most independent colleges in the state. Dr. Walters resides in Columbus, Ohio.