Re: Types of numerals; bases in natlangs.

Quoting Herman Miller <hmiller@...>:

> veritosproject@gmail.com wrote:
> > MiB, KiB are "mibibyte" and "kibibyte".  Technically, unlike what you
> > have heard, a kilobyte and megabyte are really exactly 1000 and
> > 1000000 units.  A kibibyte is 1024 bytes and a mibibyte is 1024
> > kibibytes.
>
> A kilobyte is never 1000 bytes; "byte" isn't an SI unit and doesn't
> follow the SI rules. A megabyte technically shouldn't be a million bytes
> either, but someone in a hard disk marketing department thought it was a
> good idea. In the context of RAM, a megabyte is always 2^20 bytes
> (1,048,576). No one in my experience ever uses the bizarre "kibibyte"
> and "mibibyte". The abbreviation "K" is generally used in place of
> "kilobyte" (which is rarely heard). "Meg" and "megabyte" are both in
> common use.
>
> There really is no logical reason to want a name for a block of 1,000 or
> 1,000,000 bytes. The confusion could easily have been avoided by
> sticking with the traditional definitions, and just saying "a million
> bytes" if for some reason you want to talk about a million bytes.

Still less is there any logical reason to use a prefix meaning 1000 to mean
1024. The real detrimental language change here was started by whatever cretin
first hit on the idea to use kilo = 1024.

"Kibibyte" might be 'bizarre', but "kilobyte" is just plain abuse of language.

                                                        Andreas

Reply