KiB MiB GiB has NOTHING to do with computers!!!!

bloodbob

Trollipop
Veteran
Okay I think I'll educate the people here who do not understands.

First of all anyone who says KiB or MiB or anything like those is an SI units they are WRONG.

The "National Institute of Standards and Technology" is doing a very wrong things and made this website which has listed these units under the category of "International System of Units (SI)" http://physics.nist.gov/cuu/Units/binary.html. If you read carefully the organisation that made up these units is "International Electrotechnical Commission" as you can see already this isn't and information technology body. Now the IEC isn't the body that controls the SI standards that is the Bureau International des Poids et Mesures ( BIMP ). If you go and consult the BIMP website and look at the prefixes page http://www.bipm.org/en/si/prefixes.html you will see there is no Ki or Mi or anything else prefix ( if your gonna argue about the related articles links please notice the "by the IEC" portion of the link title ).

OFF TOPIC why shouldn't a kilobyte be 1000 bytes since kilo is the SI prefix for 1000? Simply because a byte isn't an SI unit so wtf in gods name are you trying to place SI rules on it?

Okay where did KiB and MiB come about.

Well electrical engineers don't care about computer they just care about signal transistions on a piece of wire ( or firbe or air ) and basicly thats is measure in Hz which is cycles per second ( 1/s) and because a single signal can have more then two possible values in some signaling system well they made its Bits per second rather then Bit cylces per second or Bit Hertz which is rather poor sounding and the cycles doesn't add any extra meaning. Since number of bits per cycle is simple a constant and doesn't factor into 99% of the equations they just stuck with the Hertz for everything which of course is 1/seconds which is a derived SI unit and follows SI rules and makes life easy for the engineer.

Now computer scientists can quite handle network speeds being done in powers of 10s because we know the engineers don't care about powers of 2 but when we talk about data sizes rather then data transfer rates those poor engineers get confused. So they decided to make up their own units so they wouldn't get mixed up all the time.

Now NIST made a website showing off what the IEC and labbed it as an SI units website and well basicly millions of idiots are now taking the word of NIST as GOD and all the idiots have decided they will just force upon all the computer scientists the enigneers units. Since the website says KiB is an SI unit and we should all be using it.

Frankly Engineers can use it all they want but when someone tells me a file size based on powers of 2 should be using XiB I will kill them as I"m not an engineer I'm either an IT person or a Chemistry person and since neither of my backgrounds have ever used this XiB standard I don't see why I should use it now.
 
I don't know what you're whining about, everyone who works with computers should know that data capacity is measured in KB, MB, GB, TB, etc. which are all base 2 when dealing with things like RAM, and base 10 when dealing with large capacity storage devices. Anyone working with throughput should know that values are represented in bits per second, or some multiple thereof (Kbps, Mbps, Gbps, Tbps, etc.) in base 10.

I studied computer engineering for two years, and completed an associate's degree in computer networking and never saw the KiB/MiB/GiB notation in any text. I doubt it will ever catch on. De Jure standards mean nothing if no one wants to follow them.
 
bloodbob said:
If you go and consult the BIMP website and look at the prefixes page http://www.bipm.org/en/si/prefixes.html you will see there is no Ki or Mi or anything else prefix ( if your gonna argue about the related articles links please notice the "by the IEC" portion of the link title ).

OFF TOPIC why shouldn't a kilobyte be 1000 bytes since kilo is the SI prefix for 1000? Simply because a byte isn't an SI unit so wtf in gods name are you trying to place SI rules on it?
You've addressed the first statement with your second one. Since the BIPM currently do not recognise the terms bits and bytes as being "official" units, they are not obliged to consider the prefixes. However, the IEC and IEEE (who have far more involvement in the world of bits and bytes than the BIPM) have at least attempted to address the confusion; of course, NIST is not a worldwide regulartory body (unlike BIPM, IEC or CIPM) and typically uses units that are not officially recognised by BIPM and therefore not classed as SI units.

Personally, and all of the UK examination boards agree with me on this one, it's easier to have a prefix having the same meaning across the board regardless as to whether a unit is an official SI one or not. It's not impossible to get 1000 or 100000 bits or bytes, so it shouldn't be an issue that kB = 1000 bytes. The fact that it is shows just how anally retentive the world of science and engineering is.
 
Yes, In my second year CS degree exams I just stated my assumption on what I ment by KB (1024 or 1000).

On questions that related to framebuffer memory sizes I used 1024 bits in a KB and for networking and HDD storage questions I used 1000 bits in a KB.
 
Yeah the only reason we are using it for HDD storage was so that those frigging greedy companies could claim they held more :/
 
Engineer's point of view:

When I use kB & MB, it almost always happens in a colloquial sense. When I say "The board has 1MB of flash", it is insignificant whether MB is 1000000 bytes or 1048576 bytes.

When actual memory ranges and addresses are referred, they are always discussed in hexadecimal anyway. In buffer sizes etc. where decimal could be meaningful, I always use plain bytes.

This is as much of an issue as that of people speaking of energy content as calories in food while actually meaning kCal's. Cal is even an obsolete unit, replaced by joule in SI-system.
 
Back
Top