Skip site navigation (1)Skip section navigation (2)
Date:      Thu, 21 Feb 2002 16:06:07 +1100
From:      Michael Wardle <michael.wardle@adacel.com>
To:        "Gary W. Swearingen" <swear@blarg.net>
Cc:        parv <parv_@yahoo.com>, Giorgos Keramidas <keramida@ceid.upatras.gr>, Wouter Van Hemel <wouter@pair.com>, doc@FreeBSD.ORG
Subject:   Re: inconsistent use of data units
Message-ID:  <3C74803F.4090004@adacel.com>
References:  <3C743707.3080505@adacel.com>	<20020221003116.GA11893@hades.hell.gr> <3C744D39.1020308@adacel.com>	<1014256250.304.66.camel@cocaine> <3C745639.8080509@adacel.com>	<20020221022225.GA12900@hades.hell.gr> <3C745D8B.9090808@adacel.com>	<20020221025358.GB2678@moo.holy.cow> <3C7464B4.70004@adacel.com> <u3adu3bgb3.du3@localhost.localdomain>

next in thread | previous in thread | raw e-mail | index | archive | help
Gary W. Swearingen wrote:
>>>and ... unless noted otherwise, in computer context, i do not ever
>>>assume 'kilo' to represent 1000 number, only 1024.
>>>
>>This is part of the problem.  kilo only ever means 1000, and to use it
>>otherwise is incorrect.  You can not steal SI prefixes and redefine them
>>how you wish.
>>
> 
> SI prefixes are defined for use on SI units, aren't they?  As far as I'm
> concerned, "byte" is a computer-industry unit with associated jargonized
> prefixes.  When you're talking about bytes or words, k, K, M, G, T, etc,
> have non-SI meanings...

I agree that "bit" and "byte" are not formal SI units, and so do not 
formally fall under SI/CIPM "jurisdiction", however the choice to use SI 
prefixes has implicitly caused comparison between these units and SI -- 
not surprisingly -- and the exact meaning of kilo now needs to be 
clarified so as to avoid confusion, and it *can not* mean 1024, as this 
directly contradicts the international standard.

Can you think of a sensible reason why kilobyte = 1000 bytes is not a 
good idea?

Would I be correct in assuming that you do not use (and therefore are 
not familiar with) SI?

> The disk drive people are just wrong;

No, they are quite correct, however they are in the minority.

I am not aware of any official body that has standardized kilobyte as 
meaning 1024 bytes.  It may have been incorrectly assumed that kilobyte 
meant 1024 bytes, as the difference was deemed insignificant, however at 
no point could a kilobyte have been correctly referred to as 1024 bytes.
(Same for megabyte, gigabyte, and so on.)

Please point to some real evidence (preferably a standard) that proves 
the assertion (1 kilobyte = 1024 bytes) correct.

> I'd prefer some FDP introduction explain this and that "B" means "byte"
> and "b" means "bit", though I wouldn't mind a requirement to use the
> full words except where the meaning is clear from the context or a note.
> 
> 
> P.S. It's unfortunate that our industry borrowed terms and used them in
> familiar contexts with strange meanings (as I hate to see happening with
> "proprietary", BTW), but it has happened and attempts to introduce "bi"
> prefixes will cause more trouble than it solves.

What is the 100% formal, standardized method for referring to 1000 
bytes?  There is none.  What is the 100% formal, standardized method for 
referring to 1024 bytes?  There is none to my knowledge.

The IEEE/IEC standard (or is it still a proposal?) is the closest thing 
existing that attempts to clarify and standardize on one terminology.  I 
believe it should be given consideration (if not, it should simply be 
accepted), as the IEEE is definately a body relevant to electronics and 
computing (the first thing that springs to mind is networking standards 
such as IEEE 802.11).  If it were the national council of skateboarding 
(a randomly generated organization!) with this system, then perhaps it 
would be sensible to ignore it, but this is the IEEE!

If we continue to use "borrowed" SI prefixes in a sense other than they 
were intended, there will continue to be confusion.  If we do what is 
IMHO the right thing, there may be confusion for a short period, however 
there will be no ambiguity (KiB can not be mistaken for something else), 
and people will soon become familiar with this standard.

Regards

-- 
MICHAEL WARDLE                |  WORK   +61-2-6024-2699
SGI Desktop & Admin Software  |  MOBILE +61-415-439-838
Adacel Technologies Limited   |  WEB    http://www.adacel.com/


To Unsubscribe: send mail to majordomo@FreeBSD.org
with "unsubscribe freebsd-doc" in the body of the message




Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?3C74803F.4090004>