Re: Using standardized SI prefixes
On Tue, Jun 12, 2007 at 02:36:55PM +0200, Magnus Holmgren wrote:
> > And what version is used is trivial to check. Oh, wait -- in this case not
> > that trivial, when using the Google test "kilobyte" is so much over the cap
> > that you need tricks like searching for "kibibite foo" and "kilobyte foo",
> > using several words to avoid bias caused by a certain term.
> > The results I got are: "kibibite" has below 0.3% use of "kilobyte".
> > With such a crushing defeat, I doubt the whole "kibibyte" crap has much of
> > a leg to stand on, regardless whether a self-imposed "Academie Anglaise"
> > says.
> I get 59 500 hits for "kibibyte" and 1.5 million hits for "kilobyte". That's
> about 4%, not 0.3%. In fact, it's sufficiently widespread to earn a place in
> dictionaries, IMHO.
That's the cap I mentioned earlier -- most searches are halted soon after
roughly 1M hits. To have a clue about the real value you need to search for
phrases which contain the words you want to compare. Repeat this a few
times to eliminate the risk of using a helper word used more by those who
have a stronger preference for one of the variants.
1KB // Microsoft corollary to Hanlon's razor:
// Never attribute to stupidity what can be
// adequately explained by malice.