[MPlayer-users] Re: I wrote a mencoder mini-guide for DVD ripping

[D Richard Felker III]

On Sat, Aug 24, 2002 at 07:25:32PM +0300, Anssi Saari wrote:
> [Automatic answer: RTFM (read DOCS, FAQ), also read DOCS/bugreports.html]
> On Sat, Aug 24, 2002 at 12:11:45PM +0200, Dominik Mierzejewski wrote:
>  
> > > 1KB = 1KibiByte = 1024 bytes
> > > 1Kb = 1Kibibit = 1024 bits
> > 
> > NO! Yuck. That was a bad idea from the beginning.
> 
> Yes, it's disgusting. It's like some kind of a mouth exercise and without
> a problem brings the term "gibbering idiot to mind...
> 
> But the problem with "kB" is now that you don't know any more what it
> means. In some software it's 1000B, like apt-get in Debian. In other
> programs it's 1024B. So, as disgusting as KiB is, I prefer being
> disgusting to being unclear.

OK quick lesson.

K with bytes should always mean 1024.

K with bits has different meaning depending on context.

K when dealing with storage/addressing units should always mean 1024.
The idea is that it's idiotic to measure addressible storage units in
anything but a power of two. And yes hard drive vendors are lying
jerks.

K when dealing with network bandwidth, flow rates, streams, etc.
(non-addressed, non-storage data) probably means 1000. There's rarely
if ever a reason to talk about such a thing in powers of two.

Granted there is still a *little* ambiguity, but it's no big deal.
We've lived with it for many many years and can live with it still. We
mathematicians accept that the human mind is capable of dealing with
context better than dealing with the massive abundance of symbols
needed to write everything with zero ambiguity. Logicians and computer
'scientists' cannot seem to grasp this. Yes it would make sense to
have a different unit identifier in a program on a dumb computer
storing and converting units. Humans are (hopefully) smart enough not
to need this. Of course reading this list, we probably often have
doubts about that.

Just my 2 cibents... (that's 1/128 dollar, not 1/100, lol!)

Rich