WHEN COMPUTERS WERE STILL A NEW IDEA, they were such a marvel that people called them “electronic brains.” It was the ability of a thermionic valve to switch an electric current on or off that became key to the development of the first generation of true computers, in the 1940s. These valves stored data in computers and performed calculations. Transistors were invented in 1947 and did the same job, but were smaller and more reliable. They heralded the second generation of computers.
THE BINARY SYSTEM
Computers store and process data by setting electronic switches in different patterns, using a coding system called binary. An electronic switch can only be in two states—“on” and “off.” In binary, the on state is number one and the off state is zero, as seen in this diagram of a binary switch. Groups of switches represent patterns of Is and Os, and these sequences in turn stand for numbers and letters. For example, 101 is the number five.
CALCULATING WITH BINARY
The valves in early computers were connected together to form electronic circuits called logic gates that made calculations using binary numbers. The sums were used to process all kinds of data. Sums look very different in binary. 1+2=3 is written as 1+10=11. All computers still use logic gates and binary numbers.
SYSTEMS WITH VALVES
In 1950, computer pioneer Alan Turing built a valve-based computer called the Pilot ACE. It used 800 valves and helped scientists solve physics problems for many years. Bigger machines used far more valves. The AN/FSQ-7 was one of the largest computers ever built. Used by the US Air Force, it contained around 55,000 valves and weighed more than 245 tons (223 metric tons). The valves in such computers produced so much heat that they burned out and stopped the computers from working every few hours.
A PIONEERING INVENTION
This is a false-color X-ray of a vacuum tube used in early radios. Early vacuum tubes, or thermionic valves, gave rise to the more complex ones used in computers. Thermionic valves can be used to stop or allow a flow of electricity, letting them act as switches. In first-generation computers, each valve switched one bit of data (see page 16 ). A series of valves stored data in the form of a sequence of Is and 0s. As a program ran, the valves would switch on or off as the numbers they held changed. The valves could control each other and be built into useful computer circuits.
Page  | Top of Article
THE NEXT GENERATION
Like valves, transistors can also be used as switches. They work by using semiconductors—materials that carry an electric current sometimes, and block it at other times. Electricity flows into the transistor through one terminal and out through the second, but only if the transistor is switched on first by sending electricity through the third terminal. Transistors did what valves could, but were smaller, faster, cheaper, more reliable, and wasted less energy than valves. Transistors led to a whole new generation of faster computers, and tiny ones are present in most modern computer circuits.
THE FIRST BUG
On September 9,1947, the Harvard Mark II computer encountered a problem and stopped working. When its operators investigated, they discovered a bug—a moth—that had gotten stuck between the computer’s relays. Ever since then, the word “bug” has been used to refer to problems in a computer, mainly to those caused by errors in computer programming. The process of checking and removing errors from computer programs is known as “debugging.”
TRANSISTORS IN ACTION
It was only when people put transistors into computers in the 1950s that the machines became reliable enough to be really useful. Soon, several large companies began using them. International Business Machines (IBM), a computer-making company, tripled in size in the 1950s. American Airlines used an IBM 7090 computer (left) to reserve plane seats, carrying out the world’s first online bookings in 1964.
Built by Control Data Corporation (CDC), the CDC 6600 was the first successful supercomputer (see pages 30 -31 ), capable of performing more than 1 million instructions in a second—today’s supercomputers can perform more than 1 quadrillion calculations in a second. This machine contained about 400,000 transistors based on the element silicon, instead of the germanium transistors that earlier computers used. More than 100 machines were sold, mostly to academic and military research laboratories. The CDC 6600 was the fastest computer in the world between 1964 and 1969.