Smart Machines

If you look at the messy side of your PC, you'll see a jumble of wires that connect the machine to the outside world. Look more closely and you'll notice that the wires disappear into two types of slots, which your PC guru will affectionately refer to as either "serial" or "parallel" ports.

In essence, a serial port allows you to communicate over a simple dual-wire connection, like that used for a telephone. A parallel port, on the other hand, uses a larger number of wires, which can be faster. But because the extra wires increase the cost of operation and invite greater interference, parallel ports are generally used only for transmitting data over short distances, such as from your PC to your printer.

Both kinds of connections were in use long before the advent of the PC. To find out which came first (albeit by a narrow margin), we have to go back more than 160 years.

At the start of the 19th century, pretty much the only thing known about electricity was that an electric current running through a copper wire produced a magnetic field. The effect had first been demonstrated by the Danish scientist Hans C. Oersted in 1820. Eleven years later, in 1831, physicist-chemist Michael Faraday, in England, and physicist Joseph Henry, in the United States, independently discovered that the reverse was also true: a magnetic field could induce an electric current in a wire.

The discoveries set off a flurry of work by both amateur and professional inventors who tried to produce electric telegraphs that could exploit the two principles.

One such inventor, William Fothergill Cooke, collaborated with British scientist Sir Charles Wheatstone to develop the first electric telegraph actually put to use. The instrument, which relied on five signal wires (the world's first parallel device), used Oersted's principle to deflect magnetic compass needles with electric currents. A small current would be sent through two of the wires at a time. The two needles at the end of the wires would, in turn, be deflected and would together point to one of 20 different letters laid out along a diamond-shaped grid. Cooke and Wheatstone's five-needle telegraph was patented in England on June 12, 1837, and installed within days on the London-Birmingham Railway line.

In the United States, Samuel Morse was working with his assistant, Alfred Vail, and physicist Joseph Henry on a different type of telegraph: a serial device. He first demonstrated the instrument on September 2, 1837, in New York but because of a protracted legal battle wasn't granted a full patent until 1854. Morse based his design not on the deflection of magnetic needles but on the workings of an electromagnet--a magnet with wire wrapped around it that could be made to attract or repel a small metal plate. As the plate was attracted or repelled in reaction to the signaling current, it emitted a series of clicks, which encoded particular words.

Morse's telegraph had two distinct advantages. First, the moving plate could be connected to a pen to leave a written record of a communication. Second, because of the device's aural component, telegraph operators didn't have to watch the transmission but could listen to it and decode the words as they arrived. Indeed, it was that sound that inspired the famous Morse code, with its sequences of short (dot) and long (dash) clicks corresponding to individual letters of the alphabet. Within a decade, Morse's design had replaced all its competitors.

Gerard J. Holzmann is a distinguished member of the technical staff at Bell Laboratories, in Murray Hill, N.J.