The word bit is simply a compression of the term binary digit, and it was originally used by mathematician John Tukey while working at Bell Labs in the second half of the forties. But what about the byte? We all know that a byte equals eight bits, but that had to come from somewhere. Professor David Brailsford of the University of Nottingham participated in a recent video for the Computerphile channel, where he explained how the byte became what we know today.
The Quest for Decimal Precision
One of the most critical aspects to consider in this exploration of the byte is that early computers were based on words (word) with a specific number of bits. This reflected, in a way, the needs of programmers of the time, who prioritized decimal precision above all else. The general desire was to obtain ten digits of precision when the average logarithmic table delivered four. Brailsford explains in the video that ten digits equal just over 33 bits, and that led to the construction of systems like EDVAC, which aimed for a minimum of 40 bits (44 in total).
IBM's 8-Bit Proposal
Scientific computing had achieved its goal by applying different strategies, but commercial computing had gained a lot of traction, which brings none other than IBM into the equation. IBM's conclusion was that those original computers were terrible at manipulating characters. For example, applying 5-bit characters (Baudot code) to a minimum word of 18 bits was a gigantic waste and a performance nightmare.
IBM's proposal was that the word stop being the minimum unit, and that each character be addressed individually, under a fixed reference of 8 bits per unit. The word byte had already been used previously to describe other arbitrary groups of bits, but it was essentially Werner Buchholz during the development of the IBM 7030 who standardized the modern byte as an 8-bit package. The best part? By ordering the hardware to interpret four of those bytes in a 32-bit word, mathematicians and engineers had almost everything they needed at a much higher speed.