Since humanity got the idea not to limit itself any longer to the usage of producing and receiving sounds for the purpose of mutual understanding, the rise of visual communication assumed large proportions. From a humble start, scratching stone or engraving leaves of plants, writers in due course proceeded in applying newer means again and again: hides of animals, gall-nut tincture, melting lead and laser radiants marked a development of increasing refinement and differentiation. Yet, when faced with the greatest step forward in communication facilities, that of electronic dataprocessing, the application of it for producing script seemed to require a start from scratch again. A method that was based on only two available tokens presented the most intractable medium mankind ever had been forced to handle. Nevertheless, the advantages offered by the speed of transfer were so great that generations of people got convinced that the elements of Morse-code, dots and dashes only, were a comfortable way of expression, and presented a real progress. Still more rigid were the demands of the computer: only 0 or 1, on or off, flip or flop, yin or yang, to be or not to be, were the values that the atoms of understanding, the bits, were permitted to assume. If nothing had happened later on, the Stone Age had returned by way of the computer.

Thus it was a brilliant invention, to be valued equally to that of Gutenberg and Coster, and equally difficult to trace historically, to group bits in combinations of fixed length, and to assign to each of these, a pattern created from 0 and 1, a fixed meaning. The start was modest. On behalf of telegraphy a 5-bit code was agreed upon in 1931 (when nobody could imagine the rise of computers). Technically speaking, this was a big progress compared with the system of Morse, the patterns could now be punched into paper tape, with a hole for 1 and none for 0. With each pattern a character could correspond. But not more than 2x2x2x2x2=32 different patterns are available. With 26 letters and 10 digits this is not enough. Thus one pattern got the role of SHIFT ON, and one that of SHIFT OFF. The meaning of the patterns as they succeed each other depends on the status of SHIFT. In this way almost twice as much characters could be represented as without the trick. A disadvantage was that a pattern did not have any longer a unique meaning. This dilemma of being unique or not of a code we shall meet many times in the following pages.

When in the fifties the first computers appeared, a coding system had to be developed for letters and digits, and if possible also for some punctuation marks and other small stuff. But there existed yet a tradition of punched cards and chain printers, where 48 characters were the accepted limit. Though a 6-bit code allows for 64, it lasted several years before that number was adopted for computer hardware and software alike. For generations of computers this continued to be the upper limit, with the result that texts were restricted to capital letters, and in fact, to English. To Scandinavians it was allowed to sacrifice some special characters to include their own extra letters.

This state of affairs was ended by the development, in the years 1964-1967, of two standards, one International Standard for 128 characters, for which 7-bit codes were sufficient, and the internal standard of IBM, called EBCDIC, based on 8-bit "bytes", but in which only 128 "bytes" had got a meaning. ISO 646, in its American version popularly known as ASCII, would acquire an increasing influence in the course of time, from the moment that more and more 8-bit computers arrived at the market. In the chapters that follow, it will be discussed in detail which developments the coding methods for characters were subjected to, and which solutions have been proposed to the problems met. Recommendations for the Netherlands' practice are given in the last chapter.

By Johan W van Wingen
Leiden, the Netherlands