SRFI 56 Binary I/O bear (17 Jun 2004 00:44 UTC)
Re: SRFI 56 Binary I/O Alex Shinn (17 Jun 2004 02:40 UTC)
Re: SRFI 56 Binary I/O tb@xxxxxx (17 Jun 2004 06:14 UTC)
Re: SRFI 56 Binary I/O bear (17 Jun 2004 09:06 UTC)
Re: SRFI 56 Binary I/O Alex Shinn (17 Jun 2004 10:17 UTC)
Re: SRFI 56 Binary I/O Shiro Kawai (17 Jun 2004 10:40 UTC)
Re: SRFI 56 Binary I/O tb@xxxxxx (17 Jun 2004 15:09 UTC)

Re: SRFI 56 Binary I/O bear 17 Jun 2004 09:06 UTC


On Wed, 16 Jun 2004, Thomas Bushnell, BSG wrote:

>bear <xxxxxx@sonic.net> writes:
>
>> 	read-byte
>>       write-byte
>>       peek-byte
>>       byte-ready?
>
>This is fine, but don't call them bytes.  Call them "octets" if that's
>what you mean; since "byte" has a different meaning (however remote
>these days) it would be well advised to use a word that means
>specifically what you want, rather than one that only mostly does.

This is true.  A byte is the distance in bits between the addresses
referenced by two consecutive addresses. An octet is 8 bits.  We have
gotten into the habit of calling octets bytes, because we have gotten
used to architectures where it's true.

Hmmm.  We are very close though, to completely losing that definition
of "byte" -- on some of the new RISC machines, the bytes (in the
technical sense) are 16 bits long;  when you add 1 to an address, you
go 16 bits rather than 8.  But the manuals for these CPU's call this
unit two bytes (apparently with a straight face) in "an attempt to
avoid confusion."

				Bear