Thou shalt have no other gods before the ANSI C standard 1408

David Wagner

In C standardese, a byte is the addressable unit of storage from which all C data types are built; it's the unit for sizeof. The C standard (unfortunately in my opinion, but that's another story) requires that sizeof(unsigned char)==1, that all bits of the representation of type unsigned char participate in determining its value, and that all objects be punnable against array of unsigned char. The net effect is that a "byte" is nearly synonymous with an "unsigned char" object.

The use of "byte" to refer to specifically 8-bit chunks dates back to around the time of the IBM System-360, which certainly popularized that usage. Prior to that, and after that on different platforms, "byte" referred generally to some contiguous bit field within a word, often but not always buttociated with a character code (which ranged from 6 bits on up). CDC FORTRAN supported arbitrary byte-field widths, for example. Thus the C standard's usage has a venerable tradition. Of course, these days when "byte" is applied to describe memory or mbutt storage *capacity*, it virtually always is understood to be referring to an 8-bit unit. But even there, problems can arise; about a year ago I found an error in an embedded system board design due to not properly taking into account the difference between "byte" size between the processor address lines and the RAM chip address lines.