Megabyte vs Mebibyte (was Re: Language and English)

From: Eric Chomko <vze2wsvr_at_verizon.net>
Date: Fri Jan 4 15:31:31 2002

"Fred Cisin (XenoSoft)" wrote:

> > > My personal beef is with the use of 1024000 for a Megabyte. I prefer
> > > 1048576, will accept 1000000, but can't stand the use of 1024000.
> > > (A PC HD floppy is 80 tracks * 2 sides * 18 sectors per track * 512 bytes
> > > per sector -- how can you get "1.44 M" from THAT?)
> On Fri, 4 Jan 2002, Ethan Dicks wrote:
> > Besides the 1000 x 1024 method, the moniker "1.44Mb" has an additional
> > marketing advantage - it's a simple multiple of 720K (which is really
> > a 1Mb raw floppy, formatted to 737,280 bytes under DOS, or 1024 x 720).
> > It's easier to conceptualize that a "1.44Mb" floppy holds twice as much
> > as a "720K" floppy. If you called it a "1.47Mb" floppy, I think there
> > would be even more confused newbies than there are now.
>
> I think that "1.4M" does just fine.
>
> For those who are too lazy to do the arithmetic:
> A "Decimal" megabyte (10^6) would be 1000000 bytes, which would give the
> disk a capacity of 1.47 M
> An "honest" megabyte (2^20) would be 1048576 bytes, which would give the
> disk a capacity of 1.40625 M
> An "IBM disk" or "sleazy" megabyte (1000 * 1024) would be 1024000, which
> is the only way to get "1.44" M.
>
> I prefer the binary based unit, I can accept the decimal based unit, but
> there just is no possible justification for the mixed one, other than "But
> IBM has always done it that way."
>

Ironic, that the block size I work with from sending synthetic satelilite test
data is
based upon 1000 transfer frames. Yep, a transfer is exactly 1024 bytes.
Therefore,
1024000 is used by us constantly. The 1000 count is arbitrary to some degree,
but the 1024 is not.

>
> It's especially egregious, because even IBM, who use 1024000 for Megabyte
> for disks, uses 1048576 for memory. Thus, an IBM megabyte of disk storage
> will not hold the content of an IBM megabyte of memory!
>

Seems like you have lots of energy on this Fred. 1000 1K blocks. Is that a
reason
to fly off the handle?

>
> > Think of your modem (presuming you have one - not everyone does these
> > days)... 28.8kbps goes to "56K" - ignoring the fact that the FCC limits
> > the ISP end's power so that you can't achive the theoretical maximum
> > speed, it's 57,600bps; but, calling it a "57.6K" modem or even a "fifty-
> > seven, six" modem isn't as catchy as "fifty-six kay".
>
> Rounding, I don't have a problem with. In fact, I think that Alan Shugart
> was brilliant to round down the capacities of his drives, thus
> significantly reducing the support issues ("but I only got 10 and a half
> usable megabytes on that 11.1492872347653257 M drive")
> But calling the disk 1.44 is NOT rounding from 1.40625, it is inflating
> the capacity with a BOGUS unit size.
>

May I suggest using a CD instead? They hold lots more data (660MB, I think).
The average person can afford a CD burner these days.

Eric

>
> > Marketing and Mathematics - not much overlap. It was one of the reasons
> > that NIST is proposing "Mebibyte" as a term for a "binary" megabyte -
> > i.e., 1024 ^ 2 as opposed to 1000 ^ 2. People abuse the terms long
> > enough that they lose precision, and maybe you have to go out and
> > invent new, non-ambigious terms. Personally, I would rather not add
> > jargon for the sake of jargon, but we'll see how far this goes. I don't
> > think it reduces the confusion much.
>
> I would really like "Mebibyte". IFF it were to have been introduced more
> than twenty years ago! They waited WAY too long.
>
> --
> Grumpy Ol' Fred cisin_at_xenosoft.com
Received on Fri Jan 04 2002 - 15:31:31 GMT

This archive was generated by hypermail 2.3.0 : Fri Oct 10 2014 - 23:34:52 BST