Hello Laine,

   LS> Odd, those are exactly the opposite of the way I've
     > always seen B vs b used when talking about bandwidth
     > or memory (I always figured the logic was something
     > like "b is smaller than B, and bits are smaller than
     > bytes, so b = bits and B = bytes").

I've seen it written both ways.  I make no claim that my
convention is "the right way", and in non-trivial documents
I define them on first use and possibly also in a glossary.
Sometimes the context provides strong clues: although I
might have written 100M/S meaning '100 Megabytes per second'
I'm more likely to talk about '1 Gigabit Ethernet' to avoid
the tenuous assumption that 10 bits = 1 byte.

   LS> (BTW, I've never seen any differentiation between "K"
     > and "k". "M" of course has to be capitalized, since
     > "m" would mean "milli").

Since k is defined as 1,000 (kg, km etc), I was taught to
use K for 1,024 (I think in S.I. K = Kelvin, but my machines
don't run that hot ;-)  A quick prod of google produced this
USA site...

      http://physics.nist.gov/cuu/Units/background.html

            ...which has some light reading on S.I.  I'm sure
there are some European ones too.  Their page on prefixes is
at...

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

   ...and has some I had never heard of (yottabytes anyone?).
Perhaps most relevant is...

        http://physics.nist.gov/cuu/Units/binary.html

     ...I'd seen some of these used but had assumed that they
were a translation artifact.  Looks like a new bookmark for
my browser!  Interestingly they seem to use B for byte and b
for bit as you and Wojciech both suggest.

Since NetBSD is in use internationally it would seem to make
sense for us to stick to a single international standard.  I
should probably try to start using IEC one (although it's a
multi-decade habit to break ;-)

   - Andy Ball.