Ubuntu Desktop Unit Consistency (LP: #369525)

[Christopher Chan]

Neal McBurnett wrote:
> On Tue, Jun 02, 2009 at 10:59:10AM +0800, Christopher Chan wrote:
>   
>> Neal McBurnett wrote:
>>     
>>> I agree.  More details and discussion are at this ifconfig bug report,
>>> which came to the same conclusion:
>>>
>>>  https://bugs.launchpad.net/ubuntu/+source/net-tools/+bug/240073
>>>   
>>>       
>> The interface speed in base10 yes. The number of bytes transferred, NO, 
>> because that is and has always been base2. You are barking up the wrong 
>> tree with regard to ifconfig's report on RX and TX bytes. Your beloved 
>> bit_rate page is only for interface speed. So a 100mbit/s interface can 
>> be reported as 12.5MB/s interface (100,000,000bits/8 = 12,500,000bytes) 
>> which is still base10 but the amount of bytes transferred has to be 
>> base2 because that is how blinking file sizes are calculated. The size 
>> of a file has always been base2 and so this nonsense of reporting disk 
>> space in base10 will only lead to discrepancies between the amount of 
>> space available and how many files you are dump on it.
>>
>> That stupid IEC standard is at complete odds with the way computers 
>> operate. I don't want to have to miscalculate just because tools started 
>> following stupidity and gave me numbers that were rounded up or down. 
>> Take this MB/Mib nonsense and stuff it. As a system administrator, I am 
>> having NONE of it.
>>     
>
> Have you read the actual references we've been providing?  Would you
> mind providing some of your own if you disagree?  This is not just the
> IEC promoting consistent use of the metric system - it is most of the
> relevant standards bodies.  The world doesn't care that some system
> admins got used to a bad idea when it was in vogue for a short while
> in the overall history of the metric system.  Users buy disks that
> list decimal multiples on the box, and are pissed when the system
> reports it as a smaller number.  There are more users who want the
> world to agree on what the prefix "M" means, than sysadmins who want
> to redefine the metric system.
>   
Too bad it took over a decade (two?) before someone tried to sort out 
that misuse of the metric system. And they still have got nowhere after 
a decade too. Looks like the computing world don't care what the rest of 
the world thinks. Typical eh?


> E.g.
>
>  http://en.wikipedia.org/wiki/Binary_prefix#Software
>
>  The binary convention is supported by standardization bodies and
>  technical organizations such as IEEE, CIPM, NIST, and
>  SAE.[4][2][5][58] The new binary prefixes have also been adopted by
>  the European Committee for Electrotechnical Standardization (CENELEC)
>  as the harmonization document HD 60027-2:2003-03.[59] This document
>  will be adopted as a European standard.[60]
>   
Yawn. Please go rap something like the UNIX definition.

> As described elsewhere on that page, with pictures of labels and
> reference, files have been described with both properly labeled
> decimal multiples, and with mislabled binary multiples over time.  The
> insanity must stop, and imagining that people will prefer a system
> where you transmit at 1 MB/s for one second and end up with .
>
> Saying that having 8 bits in a byte affects these arguments makes no
> sense to me.  I bet most users and consumers don't even know how many
> bits are in a byte, and would see no reason to change what the
> prefixes mean based on it.
>
>   
Likewise, just pointing out these bodies makes no sense to me. Get this 
into the POSIX standard and then I'd be happy as a fish in water. Except 
for the part where I have to talk like a frog. Gribbit.