Note - this page was written primarily as a presentation for a young girl scout troop that was earning a computer-related badge.
When you look at all the things that a computer can do, from video games to controlling the space shuttle, it seems like the computer is such a complicated device that no human could possibly understand it. That turns out not to be true. I like to think of a computer as being kind of like a structure made out of Legos.
A number of years ago, I saw a huge model of a medieval village, made entirely out of Legos. When seen from a distance, it was mind boggling - how could anybody make something that huge and complicated? But when I got close up to it, I saw that it was really quite simple. There are only about 15 different kinds of Lego blocks, and each one is easy to understand. It took about 14 blocks to make one of the village people - when I saw how it was done, I knew I could make one of them easily. Once you know how to make one, it's just a matter of time (and patience) to make a hundred little people for the village. The village buildings were also simple - four walls and a roof. Each big structure could be broken down into smaller sub-structures; each sub-structure could be broken down into individual blocks. And the blocks were easy.
A computer is the same way - it is made of billions of very simple components, each of which is about as complicated as a light switch. These switches can be connected in a variety of ways to make more sophisticated structures - structures that can add, subtract, multiply, and divide, structures that can remember thousands of numbers or characters of text, structures that can move numbers and text between all of the other kinds of structures. Put a few million of these structures together, and you have a computer. But it is all based on a simple switch.
When computers were first invented, the switches were literally like light switches - there were two pieces of metal which would touch each other to complete a circuit, or would pull apart to break the circuit. A coil of wire would move the pieces of metal back and forth, connecting or disconnecting them as needed. This kind of switch is called a "relay", and the first computers had hundreds of relays. Grace Hopper did her early work on this kind of computer.
It was Ms. Hopper who invented the term "bug" to mean a problem in a computer. One day the computer she was using had a problem. They traced it down to one relay that wasn't working right. They found a moth whose wings had gotten stuck between the two pieces of metal, keeping them apart. Thus, the relay wouldn't complete the circuit when it was supposed to, and the computer didn't work right. We still call computer problems "bugs", and programmers are always talking about "stomping bugs" and being "bitten by bugs".
These relay-based computers represented the first generation of computers. However, they were very slow - relays can only turn on and off a few hundred times a second. Also they were unreliable - relays were always getting dirty and breaking down. A new kind of switch was needed.
In the 1940s, they started making computers out of vacuum tubes. I doubt that any of you have ever seen a vacuum tube - they used to be inside of radios and T.V.s. They did basically the same thing as relays, except they didn't have any moving parts. No little pieces of metal to break or get dirty. And they worked much faster than a relay - a vacuum tube can turn on and off thousands of times a second. Computers were made with thousands of vacuum tubes. With these second generation computers, programs could be more sophisticated, and computers could do many more things.
But even vacuum tubes were unreliable. Each tube contained a heating element which could burn out (like a light bulb burns out). Also, sometimes they developed leaks and air got inside them, which prevented them from working right. A large computer would typically only work for a few weeks before a bunch of tubes had to be replaced.
In the late 1950s, a new kind of switch was used in computers, called the transistor. A transistor required no heating element and no vacuum. They were also much smaller than tubes, and required much less power. They also switched faster - instead of switching on and off a few thousand times a second like vacuum tubes, they could go a few hundred thousand times a second. These third generation computers were the first to really became a huge commercial success. Instead of thousands of tubes, computers were made of tens of thousands, or even hundreds of thousands of transistors. But remember that one transistor is basically just a switch, just like a vacuums tube or a relay. Just smaller, cheaper, and more reliable.
In the early 60s, engineers figured out how to pack many transistors into one component, called an integrated circuit ("i.c." for short, or "chip"). The first i.c.s contained just a handful of transistors, but it didn't take long for the transistor count in an i.c. to grow. Modern day memory chips can have more than 8 million transistors on a square chip less than a quarter inch on a side. When i.c.-based computers first were made, the central processing unit (CPU) used to be as large as a refrigerator; now the whole thing is on a chip less than an inch on a side. Instead of hundreds of thousands of individual transistors, modern computers contain billions of them, on hundreds of i.c.s. Since the transistors are smaller and closer together, they can also turn on and off faster - tens of millions times a second. For example, if you see a computer rated at 75 megahertz (MHz), that means that the switches inside can turn on and off 75 million times in a second. These fourth generation computers are still basically the same as their ancestors, but are much smaller, cheaper, more reliable, and use less power. But it all boils down to a simple switch.
What's next? What will the fifth generation be like? Nobody knows for sure. The past thirty years have been spent making i.c.s with more and more transistors. There is still a little ways to go before we exhaust i.c. technology. But eventually we will - we will reach a point where we can't fit any more transistors on one chip. At that point, a new kind of switch will be needed. Maybe it will use lasers, or maybe super conductors. Maybe one of you will invent a totally new kind of switch that will make today's i.c.s look like clunky old relays.
