pdbain@bnr-vpa.UUCP (12/05/86)
Contrary to popular belief, a byte is not necessarily 8 bits - it just frequently happens to be that size. First some history. The infamous IBM 7030 (late 1950's, early 60's), also known as "Stretch", was IBM's first attempt at a supercomputer. Among other spiffy features, it allowed the programmer to specify the size of the operands, so if he wanted to do 23 bit arithmetic, he set the "byte size" field of the instruction to 23. The designers coined the term "byte" to mean "a bunch of bits operated on by an instruction". The 360 changed the meaning to 8 bits. Machines with 36 bits, such as the PDP-10, Honeywell mainframes, and Univac (sorry, Sperry. uh, what are they called today?) have either 6 or 9 bit bytes, and the BBN C-70 has a ten bit byte. Further, the information coding people talk about "bytes" to mean "a few bits", and talk quite freely about byte sizes. Finally, the IEEE dictionary defines a byte asa small number of bits which can be operated on by a machine. -peter