🚀 go-pugleaf

[Note: followups redirected to comp.std.internat only]

[I tried to send this earlier, but an overflowing /usr/tmp caused my file
to get eaten -- this is from memory and not as good as the original.]

	This all started when I noticed that <<octet>> is the French
translation for the English "byte", and pointed out that this is sure to
cause confusion when you teach people that a byte does not have to be 8
bits.  (This, by-the-way, was not intended as a slur against the French
language, simply an observation; English has many, greater, stupidities.)
Random people have objected to my contention that "byte != 8 bits".
Several people have pointed that <<multiplet>> is a {better, different,
alternative, what-have-you} word for <<octet>>.  I like <<multiplet>>
better for the generalized concept of "smallest natural grouping of bits on
a particular machine".

In article <1257@ncr-sd.UUCP> lodman@ncr-sd.UUCP (Michael Lodman) writes:
> Although at one time [a byte is 8 bits] might have been a correct
> statement, its my quess that a reference today to anything but an 8 bit
> byte will cause blank stares and massive confusion. For all practical
> purposes the definition of a byte is eight bits.

	If you speak Latin to most people in the world today, you will get
blank stares and massive confusion.  Does that mean Latin is not a real
language?  If I told my mother (an intelligent and educated women who
doesn't happen to grok complex algebra) that "sqrt (-1) = exp (j*pi/2)"
that would also cause blank stares and massive confusion.  Does that mean
it isn't a true statement?  Try telling all those people who work with
DEC-20's that a byte is 8 bits, and see how many blank stares you get then.

	Being common usage doesn't make something right.  I commonly talk
about making a Xerox copy, when in reality we only have Kodak and Cannon
photocopy machines at work.  When I refer to the RS-232 port on a Mac-plus,
but I know full well it's not really RS-232 because (among other things) it
has the wrong connector.  That people are sloppy about the phrase "Xerox
copy" doesn't cause much trouble (to me anyway; Xerox folks probably feel
differently), but people being sloppy about what they mean by "RS-232"
causes me more grief than you want to know about.

	It is important to make a distinction between common speech and
speaking technicly.  "Byte" is a technical term, and as such, it has a
specific meaning.  I'm not going to stick my neck out and say exactly what
the definition is, but it sure doesn't include having to be 8 bits.  If
people continue to use technical terms for whatever they feel like meaning
that particular day, pretty soon we'll have:

	If it signals with +/- 10 volts, it's RS-232.

	If it has 2 N connectors and a DB-15, it's Ethernet.

	If it's 1200 baud full-duplex, it's Bell-212A.

	If it's got "From:" and "To:" lines in the header, it's RFC-822.

	If sending it "ESC [ <something>" makes it jump, it's X3.64.

	If it prints out a "% " and runs "ls -l", it's Unix.

	Unfortunately, there are all too many people in the world who would
say that some or all of the above are true.  Many of those people are in
the business of building and selling computer systems.  More the pity.
-- 
Roy Smith, {allegra,cmcl2,philabs}!phri!roy
System Administrator, Public Health Research Institute
455 First Avenue, New York, NY 10016

"you can't spell deoxyribonucleic without unix!"