The Antikythera mechanism was a very early mechanical computer of some sort, probably designed as a device to help predict the positions of heavenly bodies.
1645 Blaise Pascal completed his 5-digit 'Pascaline', a machine that could add. While it was fairly simple, it was still one of the first calculating machines.
In 1673 Gottfried Leibniz invented a machine, based on Blaise Pascal's, but this one could also multiply and divide. In the same year, he invented his calculus.
In a letter written to the French Academy of Sciences in 1701 Gottfried Leibniz outlined and described the binary (base 2) system used by all modern computers.
The first form of automatic control was a governor on a steam engine, which generated feedback and so maintained the engine's speed at a constant rate.
In 1930, Vannevar Bush built a partly electronic computer that was capable of solving differential equations, but this did not lead to an immediate revolution.
Information may be stored on a Hollerith card (the so-called 'IBM card') in the form of holes which may be interpreted as numeric values or strings of text.
In 1975, Bill Gates and Paul Allen founded Microsoft, getting their first real start when they began producing BASIC interpreters for 6502-chip machines.
In 1979, personal computers were sold in retail outlets for the first time, and the first Sony Walkman was sold, flagging two major changes in our lives.
At the very lowest level, computers are made of combinations of AND gates and OR gates. Any operation we see is the result of these being combined by software.
In principle, there is no reason why computers should use electronics: a number of workers in the past made successful mechanical and fluid logic computers.
Computer operations are often under the control of algorithms, sets of steps to be taken that will, when followed, produce the needed answer in some form.
Many computing processes rely on some form of Boolean algebra, although this is normally hidden deep within the process, where the user cannot see it.
Computers do very few simple things, but combine these well, to do complex things. Mainly, they perform ordinary addition, subtraction, comparison and sorting.
Multiplication in a computer is achieved by carrying out many separate additions until the product is reached, often with some clever shortcuts thrown in.
Division in a computer is many separate subtractions, taking the divisor away repeatedly, counting with some clever shortcuts, the steps required to reach zero.
Numbers may be expressed in different bases. We are most familiar with base-10 or decimal, although the binary, octal and hexadecimal systems are also common.
Only in one number base, base-13, is it correct to write 6x9=42, although this is probably the most insignificant splat in this entire piece of work.
The bubble sort is an example of an algorithm that may be used in computers to sort a set of data into some kind of numeric or alphabetical order.
Information is created when data are ordered and sorted in some useful way. Statistics are used best when they summarize large data sets for quick analysis.
Some aspects of real-world computing rely on game theory to provide the most satisfactory algorithm to produce the strategy that will get the best results.
A number of computers may be linked together for distributed processing in a variety of ways, increasing their collective problem-solving power and speed.
Computers of a sort form control systems in household appliances like washing machines and sewing machines, VCRs, DVDs, cars, and in many other places.
Virtual reality makes a computer user feel, see and hear data by producing analogues of the output from a piece of analysis and using analogues as input.
The study of control systems is sometimes called robotics, even though it does not normally involve the construction of a classic humanoid 'robot'.
Wearable computers are likely to end the dominance of keyboard and mouse, using extra processing power to drive speech recognition and better visual displays.
Moore's law limits future prospects for computing: it describes the rate at which computing power increases as chips become more densely packed.
The limitation imposed by Moore's law is that there is a clear lower limit to the size of chip units that will stop chip density increasing forever.
Quantum computing may get us around Moore's law for a while by producing computational units which are much smaller and much more densely packed onto a chip.
Neural networks give computers the chance to 'learn', but this is by no means the same thing as artificial intelligence, because there is no freedom involved.
Artificial intelligence is still some distance away, and existing 'smart' programs are still a long way from showing any form of intelligence.
In 1950, Alan Turing proposed the Turing test criterion for an intelligent machine, generally still regarded as the test any artificial intelligence must pass.
While computers have the power and software to generate apparently artistic work in music and art, they cannot yet compete with human creativity.
Computers can be used to produce electronic music, and given sufficient equipment and power, have the potential to completely replace an orchestra.
Biochips and animal-machine interfaces should be a reality by around 2025, according to the best available guesses, which is not saying a lot.
In medicine, computerised tomography is important, because it takes vast amounts of data from X-rays, and produces a visual image that may be interpreted.
The information gathered in genomics work can be applied through bioinformatics, which involves carrying out complex processing of raw data to gain information.
This file is http://members.ozemail.com.au/~macinnis/scifun/splatcomp.htm, first created on February 16, 2008. Last recorded revision (well I get lazy and forget sometimes!) was on February 16, 2008.