1 |
On Sun, Mar 30, 2014 at 11:35 AM, Ciaran McCreesh |
2 |
<ciaran.mccreesh@××××××××××.com> wrote: |
3 |
> On Sun, 30 Mar 2014 05:23:08 -0400 |
4 |
> Rich Freeman <rich0@g.o> wrote: |
5 |
>> To think that counting is the one situation where it is actually |
6 |
>> possible to have perfect precision, and we still manage to mess up the |
7 |
>> units... |
8 |
> |
9 |
> Because the difference between 1000 and 1024 in this context is |
10 |
> completely irrelevant. This is the ultimate bikeshed, and ISO/IEC is a |
11 |
> rather unpleasant shade of mauve. |
12 |
|
13 |
The whole point of SI is that it works beautifully when you start |
14 |
mixing units. |
15 |
|
16 |
Storage density could be expressed in GB/mm^2 or GiB/mm^2. The former |
17 |
is FAR more useful. |
18 |
|
19 |
1 GB/mm^2 = 1PB/m^2 = 1 kB/ìm^2. |
20 |
1 GiB/mm^2 ~= 0.95 PiB/m^2 ~= 1.05 KiB/ìm^2 |
21 |
|
22 |
And heaven forbid you just use GB to mean 2^30 bytes - lots of |
23 |
opportunity for error. However, even with unambiguous units the fact |
24 |
that different elements of the compound units end up converting with |
25 |
different bases makes the math not work out cleanly. |
26 |
|
27 |
As long as you stay trapped in your nicely-discrete world of the |
28 |
computer unit complexity doesn't matter so much as long as you stay |
29 |
consistent (which isn't made easy by the fact that there are warring |
30 |
factions over the definition of the kB). As soon as you start getting |
31 |
into anything that involves the real world and engineering you can't |
32 |
avoid the units used to measure things in the real world. |
33 |
|
34 |
If you don't care about the math coming out cleanly then we might as |
35 |
well go back to using short tons, inches, and torrs. |
36 |
|
37 |
Rich |