Interesting information about the history of computing can be found at the UCD Computer Science Club web site at: CS Club home page (http://wwwcsif.cs.ucdavis.edu/~csclub/) and at the CS Club's Computer Museum homepage at: Computer Museum homepage (http://wwwcsif.cs.ucdavis.edu/~csclub/museum/homepage.html)
Why use binary arithmetic instead of base 10? Contrast the difficulty in building and using the difference engine (a mechanical machine that used decimal numbers) with the first computer built of vacuum tubes and relays using binary numbers.
Single-purpose vs. general purpose machines. Automobiles make very poor boats (except in James Bond movies). The computer is a general purpose machine. Once software is loaded, it can do various tasks that were not even dreamed of by its makers.
We saw a good videotape on the History of Computing during the last lecture with a list of important concepts/people/events to look for.
The material presented in this lecture may be found in varying forms in many introductory texts, including the image bank on the history of computing in your computers text. The series of videotapes: The Machine that Changed the World, is available in the playback review room, 1101 Hart Hall (Just inside the southeast door of Hart Hall). You are encouraged to review one or more of those tapes to learn more about the history of computing.
We are living on the fast side of a profound societal change in the rate of accumulation of information. Prior to the advent of computers, information accumulated slowly, technology was limited in the means available to deliver information (tablets to scrolls to vellum parchments to hand-set printing to typeset printing), and only with the advent of computers, television, xerography, and other image and text capture and reproduction methods has it become possible to generate the exponentially expanding volume of information that characterizes the world today. Although the computer was not solely responsible, it has played a central role in this process. How fast have computers evolved, and why?
There have been many different counting aids used by humans, beginning with fingers (the reason for our decimal system), moving to sticks and pebbles, and gradually becoming more sophisticated as time went on. Our numbering system also evolved slowly. The concept of decimal numbers being expressed in place value notation, where the number 391 really means 3 times 100 plus 9 times 10 plus 1, did not come to us in the western world until after 1000 AD, having been invented in India in about 750 and transferred through Persia to the Arab world (which is why we call them Arabic numbers). The abacus, used by East Asian cultures, did not appear until after 1500. A few early calculating machines have been found; a greek calculating machine dating from around 200 BC was recovered from the floor of the Aegean Sea. Pascal invented a calculating machine in the 17th century, and there are other examples. However, these machines all required mechanical manual input, and were lacking a fundamental requirement that is embedded in today's computers.
The first really innovative concept that is a part of the digital computer is found in the Jacquard Loom, a device that automated the setting of threads in weaving looms. The Jacquard Loom was "programmed" by wooden planks, which had rows of holes corresponding to the threads that were to be raised/lowered for each step of the weaving process. This is the concept of a program. It means that a weaving machine can be made "general purpose," in that the variety of patterns that can be created is virtually unlimited. It also led to riots by people who were afraid that the automation created by this invention would displace people from jobs. Studies have shown, however, that in this case, more jobs were created because the industries that installed these machines were able to generate more fabric and compete more effectively with other factories, creating jobs for people to handle the increased production.
This concept of a program, together with the concept of general purpose systems, is sometimes taken to be the first step toward creation of computers as we know them today. The concept was further enhanced by Charles Babbage about two decades later (1830 or so), when he designed and tried to build an automatic calculating machine called a Difference Engine, later followed by a much more creative design of a machine called the Analytical Engine. In these designs, Babbage proposed the concept of a stored program, a set of instructions that could be stored inside the computer together with data on which the instructions would operate. Unfortunately, the technology of the mid 1800s was not sufficiently precise to allow either of these machines to be satisfactorily constructed, and the concepts were largely forgotten for over 100 years. Although our ability to machine very precise instruments has greatly improved in the last century, it is unlikely that a purely mechanical device could ever have achieved anything like the success that even the first electronic computers accomplished in the World War II period.
The videotape The Machine that Changed the World: Giant Brains describes a number of interesting points about early computers. In particular its coverage of the German engineer Konrad Zuse, whose contribution was not recognized until recently, is covered well in that video and not in most textbooks. However, the descriptions of the history of computers in Capron is very good in talking about some of the other personalities and devices involved in the early days of computers. A few points that might be worth remembering:
* The first "generation" of computers was based on vacuum tubes (to provide 0 and 1 information) and telephone relay switches (to make connections between different parts of the computer. They could add at roughly the rate of one operation per second, but they added large numbers (23 decimal digits). During this generation, the concepts of stored program and stored data became increasingly fixed as a part of computer technology. The speed of vacuum tube systems gradually increased as systems became commercially available. IBM entered the business in the mid 1950s, several years after the first commercial computers were marketed by UNIVAC. However, because of their vast experience with tab product systems, IBM became the dominant force in computing within five years and continues to hold a prominent position today.
* The second generation of computers, based on transistors, were able to perform calculations at the rate of 1,000 additions per second, 3 orders of magnitude (1000 times) faster than earlier systems. These computers used punched cards for input and line printers for output, working in batch mode to process a series of jobs, one at a time. There were practically no interactive terminals available then.
* The third generation of computers was based on the use of integrated circuits and semiconductors. These systems could perform roughly a million additions per second, and they got even faster than that over the life of this technology, which is still in effect today.
* There is little agreement over "fourth generation" and "fifth generation" of computers. We have seen a continuous improvement in speed, cost/performance ratio and use, and many people have their own definitions of these boundaries.
In our review today we saw examples of the evolution of several computer components. Beginning with a two-vacuum tube "module," we illustrated a 10-transistor module, and then a simple integrated circuit. We also showed the changes that have occurred in going to successively smaller integrated circuitry, from IC (simple integrated circuit) to LSI (Large-Scale Integration, to VLSI (Very Large Scale Integration), and on to VVLSI (very very large-scale integration. The size of the chips has grown as manufacturers developed techniques to take advantage of better production techniques to create larger chips. In addition, the packing of components on each chip has greatly increased (almost to the finest limits available with laser etching of the silicon wafers used). Examples of chips shown in class included the Intel family: first the 8080, then the 80286, the 80386 (usually called the 386), the much larger 486, and currently the Pentium (P5) and Pentium Pro (P6). We also showed examples of the Motorola 68000 and its successor the Motorola 68020.
There are many different viewpoints from which we can study the history of computing. One such viewpoint relates to the size of the computer, in terms of its sheer bulk and also its cost. To consider computers in this way, it is instructive to draw an analogy between mankind's use of energy for work and the ways that we have made use of different modes of energy.
For the first several thousand years of the evolution of mankind, the only convenient form of energy available to perform tasks greater than could be performed by a single person was the work that could be extracted from living beings: wives, horses, oxen, slaves - whatever means could be devised to harness these sometimes fractious sources of energy were used at one time or another to get major jobs done: building large structures, harvesting crops, grinding grain, or laying siege to a neighbor's castle. As a result, not much in the way of very large works was accomplished. True, some pharaohs, by enslaving a few thousand captives, managed to construct some rather impressive pyramids, and a few other grandiose schemes were accomplished (such as the statues on Easter Island), but for the most part, the sources of external work to accomplish the dreams of mankind were extremely limited.
With the discovery of water power, a gradual but profound change took place. As people realized that they could harness the power of running rivers, mills came into being. Areas where water power was easy to harness became centers of large scale work. With the advent of the industrial revolution in the 19th century, this effect became much more profound; areas such as the "Fall Line" in New England, noted for abundant cheap sources of water power, came to dominate the textile and other industries spawned by the industrial revolution. Energy for mechanical work was concentrated in those areas that could harness it cheaply, and other parts of society came to depend on these central sources of power for goods that were essential to the economy of that period.
With the advent of steam, and then of electricity, another profound change took place. Steam transformed the world by providing a portable form of power. In fact, it led to the steam engine and to trains capable of transporting vastly more goods than could be carried by manual methods that it replaced. Steam was also used to power some factories, but an even more important change occurred when man discovered how to harness electrical power. This discovery led to the ability to "distribute" energy across wires throughout large areas. As a result, the centralized power structure of pre-electricity days was threatened. Mill-towns lost their dominance, and industry was able to relocate wherever the basic goods used in the manufacture of a product were the most readily accessible. This process has continued to the present day. As it became easier to construct smaller electric motors and to build more powerful and longer life batteries, motors became more widespread, smaller, and more "invisible" in the manufactured products used throughout the world. Today, we don't know (or care) how many motors are in our homes (kitchens, automobiles, toys, appliances). We take them for granted. When there is a general power failure, we may be more aware of the shortage, but for the most part we accept the widespread availability of electric motors as a given in our society.
This historic review bears a strong resemblance to the evolution of computers. In the 1940s, there were a half-dozen computers, designed in clumsy ways, using expensive (vacuum tube) components, consuming vast amounts of power, and taking up large amounts of floor space. No one in those days thought that more than a few dozen such machines would be required in the world. This was the era of the mainframe, a large, expensive, central processing unit designed to process one job at a time. The computers of the mainframe era cost a million dollars or more each. Because they were so expensive, they were in a central location: one such computer would serve the needs of an entire government agency; a large corporation would have a single computer in its main office. Mainframes were managed by data-processing professionals - individuals who had grown up with the tab product generation of business machines that were in widespread use in the first half of this century.
Tab Product systems raise another important historical connection to the 19th century. In the late 1880s, Herman Hollerith, then an employee of the United States Census Bureau, devised a new mechanical means for collecting and analyzing the formidable amount of data that was required to gather information about every person living in the U.S.A. at the time of the 1890 census. He invented the punched card, sometimes referred to as the Hollerith card, or more informally as the IBM card. The Hollerith card had 80 columns and could represent 80 characters (numbers or capital letters). It served the 1890 census well, and it served Mr. Hollerith well, for he left the Census Bureau and formed a company that later became known as International Business Machines, Inc., using punched cards and associated equipment to tabulate products and data for corporations from the early 1900s until these systems were replaced by the computers of the 1950s. The punched card remained an important part of the mainframe era; almost all input to computers was based on punched cards (data as well as programs). Keypunch operators became an important part of data processing shops, and programs were designed based on the assumption of 80-column input records. We retain this heritage, since almost all screens and terminals today have 80-column displays, and many data records are governed by an 80-column limit to the size of the record.
Mainframes dominated the computing world from 1950 through the early 1970s. During the 1960s, however, with the technological advances made possible by large scale integrated circuits, a new type of computer evolved: the minicomputer. Minicomputers were smaller, cheaper, and simpler. Their operating systems were often designed to be special purpose rather than general purpose. One important area in which the minicomputer began to play an important role was in support of laboratory systems: input, storage, and analysis of data generated by laboratory instruments. One of the first of these systems was the LINC, standing for Laboratory Instrument Computer. Originally designed and built at MIT's Lincoln Laboratories (one of the early sites of mainframe computers), the LINC evolved into a commercial product marketed by a fledgling corporation named Digital Equipment Corporation, which grew to become the second largest computer manufacturer in the world (after IBM) based on its success in the minicomputer field. (A faculty member in the School of Medicine participated in building one of the early LINC computers, and he brought his system to Davis in the early 1970s, using it for many years in his research in medical pharmacology.)
The effect of the minicomputer was somewhat analogous to the introduction of steam power in the industrial revolution. It made possible the use of computers by groups who could not have afforded to buy mainframes. Although minicomputers were relatively limited in power and fairly expensive (the cheapest probably cost somewhere around $40,000 in their early days), they broke the stranglehold of mainframes on computing applications in government, industry, and especially in universities. Prior to the availability of minicomputers, researchers had no way of harnessing computers to analyze research data. Mainframes could not be linked directly to laboratory instruments, and the data were too voluminous to keypunch manually. Mainframes also cost too much to be affordable by departments or small companies. With the advent of minicomputers, it became possible for research units to acquire minicomputers, encode laboratory data, and embark on scientific applications of computers that had previously been almost unthinkable.
The changes introduced by the minicomputer continued. Large-scale Integrated circuits (LSI) gave way to Very Large-Scale Integration (VLSI), and the concept of embedding a few thousand transistor-like components on a single chip became feasible and then became widespread. As the number of transistor-like components available on a single chip grew, it became feasible to consider putting an entire Central Processing Unit (CPU) on a single chip.
One of the first really successful CPU chips was the Intel 8080 CPU. This was an 8-bit (one byte-based) microcomputer. It led to the development of microcomputers, which first appeared in about 1975. (Your instructor, Dick Walters, built his first microcomputers from kits based on the 8080 during a sabbatical in 1976.) The Microcomputers of these days were somewhat limited in power, but we managed to convert the UCD Human Performance Laboratory's minicomputer- based system to one that worked with microcomputers as early as Fall, 1976.
As time went on, these systems grew more powerful. By 1980, Intel (and other computer chip manufacturers) had developed 16-bit microcomputers. In 1981, IBM introduced the concept of the Personal Computer, and the revolution that you are now familiar with got its start. The PC (Personal Computer) era derived from IBM's entry into the field, even though many of us were already using Microcomputers before that time. Gradually, these machines began to replace first the minicomputer, and in more recent times, the mainframe. Today, it seems likely that both minicomputers and mainframes will disappear, giving way to networked and clustered PCs and workstations. The effect of this change will be an important one for you to watch in the years ahead.
Go to the index of
lectures for ECS15 - Fall 1997 .
Go to the homepage for ECS15 - Fall 1997 .