[ On Friday, February 21, 2003 at 00:09:49 (-0500), James K. Lowden wrote: ]
> Subject: Re: du(1) with gigabyte option.
>
> It would be nice if du and friends would report decimal values instead of
> power-of-two approximations.  "kilobyte" has a fixed meaning: 1000 bytes. 

Oh, boy.  Did you ever step in it!  :-)

A "kilo_byte_" has ALWAYS been 1024 bytes, at least as far as I can
remember.

A "byte" is not an SI unit, so why the heck do you think a kilobyte
would/should be one?

If you want to express your measurements in SI notation then you should
be counting bits, not bytes.


> http://physics.nist.gov/cuu/Units/prefixes.html

Computers and computer storage units are not measured by physicists.


> http://www-cs-staff.stanford.edu/~knuth/news99.html (see "What is a
> kilobyte?")

Irrelevant.  :-)

To possibly mis-quote somone else who's written on these matters before:

	The industry is confusing enough; we don't need to complicate
	matters more.

which is ecxatly what this lame (but somewhat humorous) attempt by the
IEEE and NIST to introduce those new silly sounding and looking terms
has done.  "byte" was bad enough and now they want to make it worse!
(how wyde is your byte anyway?)

Here is a much more sane view of these things (IMHO):

	http://www.magictree.com/dataprefixes.htm

Besides, the only problem is with the abbreviations -- by now practially
everyone in the whole world (except disk drive salesmen) knows that if
it's got "byte" in it then it means some power of two.

While I like the analogy with RAM measurements (which are by design
likely to be built and sold in units which are powers of two bytes for
the foreseeable future), this short quote by "Malor" from the slashdot
discussion a couple of years ago sums it up well too:

     Anyone will understand 'marketing gigabytes' versus 'real
     gigabytes'.  This usage makes the original lie obvious, and will
     help to correct the problem by gently reclaiming the correct word,
     instead of forcing a new one down people's throats.

For that matter most disk accesses are, and will be for the foreseeable
future, also done in blocks that are measured in units of some power of
two bytes as well.

-- 
								Greg A. Woods

+1 416 218-0098;            <g.a.woods@ieee.org>;           <woods@robohack.ca>
Planix, Inc. <woods@planix.com>; VE3TCP; Secrets of the Weird <woods@weird.com>