Home Up
Home Teaching Glossary ARM Processors Supplements Prof issues About

What is a Computer?

It’s very easy to use the term computer glibly without bothering to say what it actually means. Part of the reason for this situation is that today’s computer is virtually synonymous with the PC. This box briefly looks at some of the ways in which we use the term computer.

1, The human calculator The computer historian David Alan Grier wrote a book called When Computers Were Human that describes a period of history when people performed scientific calculations by hand. In the Victorian era, some people were professional computers and spent their working lives doing arithmetic with a pencil and paper.

2. The mechanical computer In order to increase the productivity of human computers, mechanical devices were invented to mechanise the operations of addition, subtraction and multiplication. Later, these basic mechanical devices were extended to perform  the more complicated calculations involving polynomials.

3. The Analog Computer  Largely forgotten today, the analog computer used electronic circuits to synthesise physical systems. For example, an analog computer can imitate the behaviour of a car’s suspension system by putting together components that mimic the behaviour of springs and dampers in a car’s suspension system. You don’t program an analog computer in the normal sense of the word program. You construct an electronic equivalent and then observe its behaviour. You do not have to write a program.

4. The electro-mechanical computer A natural extension of the mechanical computer is the electro-mechanical computer that uses electricity to control the moving parts of a mechanical computer. Electro-mechanical switches (relays) were used to build digital computers in the 1940s. These had some of the characteristics of today’s electronic computers.

5. First generation electronic computers The first electronic computers used vacuum tubes rather than electro-mechanical devices. These were hardly different to the human computers or the electro-mechanical computer; they were merely faster. They could not be programmed to solve different problems.

6. The von Neumann machine The so-called von Neumann machine provides the template for what we normally call a computer today. It is a device that can be programmed and its program and data are located in memory.

7. The dedicated computer The dedicated computer is designed to solve a single problem or set of problems; for example, the computer in an automobile’s anti-skid braking system.

8. The general-purpose digital computer When most people use the term computer they are referring to the general-purpose stored-program binary von Neumann machine that is exemplified by the PC. A general-purpose computer can be programmed.

9. The mainframe Once, the mainframe computer was an immensely expensive general-purpose computer that had its own building and a team of people to run it. Today, the average desk-top computer is more powerful than a mainframe of a few years ago. Mainframes still exist in very specialized applications and are invariably based on the use of multiple computers operating in parallel.

10. The microprocessor The microprocessor is a computer on a chip. It can be part of a general-purpose computer or part of a dedicated computer. Countless low-cost microprocessors are used as dedicated computers to replace electronic circuits in applications such as microwave  cooker controllers. Such microprocessors usually have sufficient memory to hold programs and data together with input/output circuits on a chip costing a few cents.

11. Embedded computer The embedded computer is a specific form of the dedicated computer that is intended to control part of a much larger system. Embedded systems often have real-time facilities; that is, they respond to external events within an appropriate time span. An embedded computer might be used to control, for example, a chemical distillation process.

12. Reconfigurable computer A significant change in computer technology occurred in the 1980s with the introduction of complex programmable logic elements; that is, the interconnection between logic elements on a chip could be performed by programming. Today we have logic systems that can be programmed and reprogrammed, and complex circuits can be modified under software control. Now, we can not only change a program, but we can also change the hardware on which it runs. This has important implications for the small-scale economic production of digital systems.

12. Quantum computer Today’s digital computers use semiconductor technology based on silicon. Quantum computers make use of the quantum states of individual atoms to store information and perform computation. In theory, quantum computers are immensely more powerful than semiconductor-based computers. The most commonly quoted potential application of quantum computers is in decrypting encoded messages. Consequently, the organizations with the strongest interest in quantum computing have three-letter mnemonics. Although, the principles of quantum computing have been demonstrated, a practical quantum computer is not yet feasible.

13. Biological computer The complex organic molecules of DNA can carry encoded information in the form of nucleotides. By performing chemical operations on these molecules, computation becomes possible. It has also been demonstrated that logic gates can be synthesized from individual cells.  Although very simple functions have been created, biological computing is still only a dream.

14. Neural computers A neural computer, or, more commonly, neural network is modelled on the behaviour of the neural cells of the brain. A neuron has several inputs and an output. The neuron fires if the weighted sum of the inputs exceeds a certain threshold. The weighted sum means that the value of each input is multiplied by a constant and all products summed (the multiplier constant is different for each input). A neural computer is created using interconnected artificial neurons. Unlike quantum and biological computers, neural networks are already a practical technology. Moreover, they can be programmed by learning (matching inputs with outputs). Neural nets are not intended to create conventional computers but to solve specific problems that are related to pattern matching; for example, reading handwriting, or predicting the movement of the stock exchange). The neural network is more related to the analog computer than the von Neumann machine.

15. Fuzzy logic In the 1970s fuzzy logic was introduced to model systems that were either difficult to quantify or uneconomic to solve using conventional computers. Fuzzy logic allows you to build a control system (for example; an automatic breaking mechanism on an underground rail carriage) with an incomplete modelling of the system. Inputs are quantized into regions or zones and then fuzzy logic operators are applied to them. For example, IF speed less than 5 AND acceleration less than 4 and distance to go greater than 10 THEN apply level 6 braking.