Mail Archives: djgpp/2000/08/28/00:33:05
Interesting... I'm keeping this handy for future reference :-)
On Sun, 27 Aug 2000 08:35:59 GMT, dontmailme AT iname DOT com (Steamer)
wrote:
>Radical NetSurfer wrote:
>
>> /* Still waiting to learn what a 16-bit entity is called on a 32-bit
>> platform (since WORD has been made in an ambiguous term. */
>
>There has never been a standard word size, so "word" has always been
>ambiguous. On a particular platform it ought to be obvious what "word"
>means, but 386-compatible processors running in 32-bit mode represent
>a problem - their native word size is 32 bits, but they are derived
>from 16-bit processors where "word" has always meant 16 bits.
>
>What's wrong with calling a 16-bit entity "two bytes"? (Or "two octets",
>if you want to be totally unambiguous.) The term "short" is OK too, as
>long as you only use it when talking about compilers on which a short is
>16 bits (which includes all x86 C compilers that I've ever seen).
>
>Or you may like to use the terminology that Donald Knuth has adopted
>for future volumes of The Art of Computer Programming:
>
> 1 byte = 8 bits
> 2 bytes = 1 wyde = 16 bits
> 2 wydes = 1 tetra = 32 bits
> 2 tetras = 1 octa = 64 bits
>
>The words "tetra" and "octa" are short forms of "tetrabyte" and "octabyte".
>
>S.
- Raw text -