metric version of 1000?

[Glenn Steen]

On 02/05/06, James Gray <james at grayonline.id.au> wrote:
> On Mon, 1 May 2006 09:36 pm, Jeff A. Earickson wrote:
> > Julian,
> >
> > > - - Support for "k", "m" and "g" multipliers in MailScanner.conf so that
> > > entries can be written as "Max SpamAssassin Size = 30k" instead of
> > > "30000". "k" = 1000, "m" = 1000000, "g" = 1000000000.
> >
> > I see you use the metric version of k, m, g.  In America we tend to use
> > the old style version of 1024 (2^10), 1048576 (2^20), and 1073741824 (2^30)
> > since we only have two fingers to count with.  :)
>
> But look at the difference:
>
> 2^10 bytes = 1024,       10^3 bytes = 1000.       Difference 2.4%
> 2^20 bytes = 1048576,    10^6 bytes = 1000000.    Difference 4.9%
> 2^30 bytes = 1073741824, 10^9 bytes = 1000000000. Difference 7.4%
>
> Granted the 1G (base 10) and 1G (base 2) difference is starting to diverge
> reasonably significantly, but once you're at the point of blocking messages
> around the gigabyte size, is blocking a message 7.4% "early" going to make a
> significant difference??
>
> I'm with Julian ;)  BTW, thanks - this mod makes the config a LOT easier for
> humans to read and manage.
>
> Cheers,
>
> James

Why are you all "upset" about this?
This is _very_ old news. (The long and short of it is actually that
the HDD makers were actually right. That they might have had another
motive than promoting a standardsbased view is... beside the point:-).

As someone mentions, there are approved ways of stipulating binary
multiples now, so... use them if you need to make the distinction
clear;-). One might argue that Jules is simply "enforcing" the SI
normative in a consistent way... Which is a very good thing.
Makes life so much less ambiguos when we don't have to have any more
"1.44M" diskettes (that are actually 1440 KiB... Sigh) and suchlike.

--
-- Glenn
email: glenn < dot > steen < at > gmail < dot > com
work: glenn < dot > steen < at > ap1 < dot > se