## 0394: "Kilobyte"

This forum is for the individual discussion thread that goes with each new comic.

Moderators: Moderators General, Prelates, Magistrates

rwald
Posts: 153
Joined: Mon Jan 29, 2007 7:14 am UTC
Contact:

### Re: "Kilobyte" Discussion

Landak wrote:I believe [http://arxiv.org/abs/quant-ph/0002077] that Qubits are represented ideally as a linear superposition of the two orthogonal basis kets |0> and |1>; i.e. the qubit \psi is |\psi> = a|0> + b|1>, where {a, b} are (complex) probability "amplitudes". Normally a normalisation convention is adapted so that |a|^2+|b|^2=1. The actual intricacies of quantum computation are incredibly complicated (and beyond my brain at the moment - I'm an undergraduate) but I heartily recommend some of DiVincenzo's more readable papers, like that earlier (it's quite understandable and doesn't really require much background knowledge) and the slightly more practical http://arxiv.org/abs/cond-mat/9911245. If you don't know what kets are, the wiki article on Dirac's bra-ket notation is somewhat useless; I suggest Dirac's "Principles of quantum mechanics" which will explain the whole lot from the very beginning. (Pracie: Generalisations of vectors)

And yes, I did register just to post this...

That reminds me of an awesome t-shirt I saw somewhere with the following written on the front:

|fu>

You keep your "10 types of people" shirts, I'll stick with shirts only 1/10 of the "10 types of people" wearers will understand.

dr7
Posts: 59
Joined: Mon Feb 12, 2007 3:17 am UTC
Location: Right here. You can't see me?
Contact:

### Re: "Kilobyte" Discussion

ARGH!! Pet Peeve! I hate -- with the fiery hatred of a thousand hating suns, you know, the kind that are in Super Mario 3? -- those -- the whole "kibibyte/bit/what the hell ever" garbage.

Look, people. I don't care about what "standard" notation for kilo/mega/etc are. This is computer world!

(At this point, I kick the messenger into the pit.)

1024 is the standard here, and if you don't like it, get out. Please.

8 bits = 1 byte.
1024 bytes = 1 kilobyte.
1024 kilobytes = 1 megabyte.
1024 megabytes = 1 gigabyte.

And so on.

Can someone please slap some sense into these people who insist on this gibi/kibi garbage, and tell them to get with the program?

dr7: king of the wicker people

Landak
Posts: 8
Joined: Mon Mar 10, 2008 8:23 am UTC

### Re: "Kilobyte" Discussion

rwald wrote:
Landak wrote:I believe [http://arxiv.org/abs/quant-ph/0002077] that Qubits are represented ideally as a linear superposition of the two orthogonal basis kets |0> and |1>; i.e. the qubit \psi is |\psi> = a|0> + b|1>, where {a, b} are (complex) probability "amplitudes". Normally a normalisation convention is adapted so that |a|^2+|b|^2=1. The actual intricacies of quantum computation are incredibly complicated (and beyond my brain at the moment - I'm an undergraduate) but I heartily recommend some of DiVincenzo's more readable papers, like that earlier (it's quite understandable and doesn't really require much background knowledge) and the slightly more practical http://arxiv.org/abs/cond-mat/9911245. If you don't know what kets are, the wiki article on Dirac's bra-ket notation is somewhat useless; I suggest Dirac's "Principles of quantum mechanics" which will explain the whole lot from the very beginning. (Pracie: Generalisations of vectors)

And yes, I did register just to post this...

That reminds me of an awesome t-shirt I saw somewhere with the following written on the front:

|fu>

You keep your "10 types of people" shirts, I'll stick with shirts only 1/10 of the "10 types of people" wearers will understand.

Lol! Cue lots of "Hey baby, wanna be my inner product so I'd create something perpendicular to the plane of both of us", and similar maths-style jokes...

grez
Posts: 1
Joined: Mon Mar 10, 2008 9:02 am UTC

### Re: "Kilobyte" Discussion

dr7 wrote:1024 is the standard here, and if you don't like it, get out. Please.

8 bits = 1 byte.
1024 bytes = 1 kilobyte.
1024 kilobytes = 1 megabyte.
1024 megabytes = 1 gigabyte.

And so on.

Can someone please slap some sense into these people who insist on this gibi/kibi garbage, and tell them to get with the program?

Unfortunately that is not consistent with SI prefixes, some of which predate "computer world" by a couple hundred years. The bottom line is that kilobyte is ambiguous whereas kibibyte isn't.

dr7
Posts: 59
Joined: Mon Feb 12, 2007 3:17 am UTC
Location: Right here. You can't see me?
Contact:

### Re: "Kilobyte" Discussion

grez wrote:Unfortunately that is not consistent with SI prefixes, some of which predate "computer world" by a couple hundred years. The bottom line is that kilobyte is ambiguous whereas kibibyte isn't.

That doesn't matter. Computers measure by powers of two, which is inconsistent with base-10. We borrow the terms because they make the most sense, but they have nothing to do with the "real-world" SI prefixes and their meanings.

Basically, if you're on a computer, you need to use computer measurements, and that ain't base 10. Otherwise, you get what we have now -- two measurements which do nothing but confuse the hell out of people. Stick with what is native to the system, and that's the 1024-based stuff.
dr7: king of the wicker people

poisonmol
Posts: 7
Joined: Tue Nov 27, 2007 9:50 am UTC
Contact:

### Re: "Kilobyte" Discussion

KiB used in Quantum Computing. -snigger snigger- That was highly amusing.

markfiend
Posts: 484
Joined: Fri Jul 06, 2007 9:59 am UTC
Location: UK (Leeds)

### Re: "Kilobyte" Discussion

Stan Kelly-Bootle is my significant other's grandmother's brother.

That is all.
Attoc dna Sublab evol eht teews secoiv fo eht slrug
pronouns: he/him/his

Stickfodder
Posts: 28
Joined: Sun Nov 18, 2007 11:36 am UTC
Location: it changes according to my mood i could be here i could be two inches to the left

### Re: "Kilobyte" Discussion

I'm even more confused now than i was before

akirjazi
Posts: 36
Joined: Wed Feb 20, 2008 1:04 pm UTC
Location: Macedonia
Contact:

### Re: "Kilobyte" Discussion

Really, how many of you care whether Jena's tits are coming at 2x10*6 bits/s or 256 kB/s or any other unit (and how much is that in Bauds)? As long as it's fast enough nobody gives a shit... btw, I wonder how long it would take, on average, to load the aforementioned tits over an average broadband line, assuming 75% perceived quality of the picture is acceptable to your personal taste... NOOO, I WAS JUST KIDDING, DON'T START CALCULATING...

MrAlex
Posts: 5
Joined: Mon Mar 10, 2008 10:50 am UTC
Contact:

### Re: Intel jokes

netsplit wrote:When Joe-average is looking for networking equipment chances are he's not gonna know the difference between bits and bytes.

And thank God we seem to have done with the term "baud"...
_____________________________________
tribù digitale blackmambo
http://www.blackmambo.it

phlip
Restorer of Worlds
Posts: 7548
Joined: Sat Sep 23, 2006 3:56 am UTC
Location: Australia
Contact:

### Re: "Kilobyte" Discussion

General rule of thumb:
If there is some technical reason why the size of the thing you're referring to should be a power of 2, then it'll usually be a binary prefix... otherwise it'll usually be a decimal prefix.

For example:
The way RAM is built, it will always be a power of 2 bits. You technically can do other values, but it'll be just as cheap to just round up to the nearest power of 2. RAM is always measured with binary prefixes... a 1GiB stick is easier to talk about than a ~1.07GB one.
CPU speed is always measured with a decimal prefix... there is no reason why it should always be a power of 2 in Hz, so there's no reason why 1kHz should be 1024 Hz.
Bandwidth, similarly, has no reason to be a power of 2... you have 10Mb/s, 100Mb/s, 1Gb/s... all arbitrary numbers, and decimal is the easiest way to choose milestones.

Hard disks are a little more complicated... namely because the disks are generally decimal prefixes, but the sizes of the files on them are generally binary prefixes. This confuses people to no end, and certainly should be changed... but it's hard to change de facto standards.
If I had to guess a reason, I'd say that it's because hard drives have never had a reason to be restricted to powers of 2... but data did, when it predated hard drives as a storage mechanism, and had to be stored directly in whatever memory was available at the time, which would've been a power of 2 in size. As a folk etymology, it works enough for me, even if it's completely wrong (though, if it is completely wrong, then I'd be happy to be corrected).
Though, most common filesystems still break up files into 4KiB blocks (or some other power of 2), so at least the "Size on Disk" measurement should be binary, at least when the value is in the KiB range (when it gets up to MB/MiB, this aspect doesn't matter so much).

Off the top of my head, I can't think of anything other than RAM sizes and file sizes that generally get measured with binary prefixes... and only the former really warrants it.

Also: Very yes to the "thank god baud is gone" motion.
Last edited by phlip on Mon Mar 10, 2008 11:42 am UTC, edited 1 time in total.

Code: Select all

enum ಠ_ಠ {°□°╰=1, °Д°╰, ಠ益ಠ╰};void ┻━┻︵​╰(ಠ_ಠ ⚠) {exit((int)⚠);}
[he/him/his]

endolith
Posts: 229
Joined: Tue Jan 01, 2008 2:14 am UTC
Location: New York, NY
Contact:

### Re: "Kilobyte" Discussion

'; DROP DATABASE;-- wrote:Hard drives are the worst for this though, where "250GB" = ~232GB. I still don't see how they get away with that.

Because that's the way that engineers and hard drive manufacturers have measured for all of eternity?

The only reason it's confusing is because Windows reports the size in powers of 1024, for some braindead reason. Why should the same partition be reported as "233 GB" in one location and "238,418 MB" in another location? What possible benefit does this convention provide? Why can't they just use the metric system like everyone else and report it as "250 GB" and "250,000 MB"? Doesn't that make a lot more sense?

aeflash wrote:It's not like we need to know that a 30 second 256kbps MP3 takes up exactly 983040 bytes = 960.000/ 983.040 KB

A 30-second 256 kbps MP3 would take up 960,000 bytes, plus any file format overhead. "kbps" is decimal.

dr7 wrote:1024 is the standard here, and if you don't like it, get out.

No it's not. Hard drive capacities, network speeds, processor speeds, DVD capacities, and a number of old floppy format capacities have always been measured in multiples of 1000.

And many of the first computers were decimal, not binary.
Last edited by endolith on Mon Mar 10, 2008 1:43 pm UTC, edited 1 time in total.

Vanguard
Posts: 807
Joined: Mon Nov 19, 2007 3:52 pm UTC
Location: Away from you

### Re: "Kilobyte" Discussion

Didn't understand a WORD of this, but I did grin at "snice you're such a good customer" but only because it felt out of place. Which was funny.

yt2005
Posts: 2
Joined: Mon Mar 10, 2008 1:24 pm UTC

### Re: "Kilobyte" Discussion

I feel like there isn't enough humor in this one to justify a text-only comic.

kriel
Posts: 922
Joined: Thu Feb 07, 2008 2:58 pm UTC
Location: Somewhere I'm not.
Contact:

### Re: "Kilobyte" Discussion

yt2005 wrote:I feel like there isn't enough humor in this one to justify a text-only comic.

It's a computer geek thing. I dunno if it was really worth throwing up on XKCD, maybe a tidbit in the blag somewhere.

Dragons Opinions be here:
But seriously, I see a lot of people in here voting for popularizing the idea of powers of 10 for computer spaces.

Here's the problem. EVERYTHING about a computer is based on powers of 2.

RAM comes, almost without exception, in powers of 2. Processors 'think' in powers of 2. Headers for files, packets, frames, etc usually use powers of 2 (or at least half-powers, ex: 4+8=12,8+16=24,16+32=48) as their number of bits. Buses are measured in powers of 2. The few things that aren't easily measured in powers of 2 are analog speeds or clocks, and guess what? We've got digital counters that work on powers of 2, to keep track of how many of those analog pulses have gone by.

If we're really going to revert back to 'True' metric, using powers of 10, knock EVERYTHING back down to bits and start using scientific notation. That'll get rid of the 'false' advertising of hard drives. (I want to fit my 1024 bytes of data into one (kilo|kibi)byte of space, damnit.) And... fix some other stuff, too. cause I ran out of material to rant on.

GCM
Posts: 528
Joined: Sat Jan 19, 2008 1:28 pm UTC
Location: Metropolis City, Planet Kerwan, Solana Galaxy
Contact:

### Re: "Kilobyte" Discussion

Somebody should do something about millibytes!! Wahahaha!

It's all binary, right? All in 1s and 0s, where a byte is 2^0=1, a kB is 2^10=1024, an mB is 2^20= 1048576, etc.

Seriously, though
DragonHawk wrote:
netsplit wrote:

No, quoting figures that aren't actually based on the technology would be deceptive advertising. The bitrate is a fair, accurate measure of the speed of the equipment. Throughput in bytes is going to vary. Do you think the manufacturer is going to quote the smaller figures for throughput? So if they did as you ask, they'd be putting a figure you'd be guaranteed never to obtain on the box. Tell me how that's better than what we have now? Joe-Average will still know that 1000 Mbit/sec is bigger than 100 Mbit/sec, which is all he needs for chest-thumping purposes.

Actually, it can work, just not in a 'comparison with other services' type way. Say they're selling a 512kb (bit)it connection. The customer will think, "Ah, 512 kBs (bytes, because thats what they're used to)". And then they think that they can download a, say, 200 Megabyte game in 200X2=400seconds=Seven minutes, when its actually going to take close to an hour.
Although, considering the sloppiness and unreliability of MY service, it actually runs slower than the 1Mbit ( Sometimes Youtube videos load at the same of faster speed than I'm watching, but I sometimes have to wait freakin long for them)

And 1024i bytes? Damn, I don't even wanna think about it.
All warfare is based on heavily-armed robotic commandos.
~Sun Tzu

Notes: My last avatar was "Vote Robot Nixon", so I'm gonna keep a list here.

Chipersoft
Posts: 24
Joined: Fri Aug 17, 2007 2:33 pm UTC

### Re: "Kilobyte" Discussion

Eleyras wrote:My favorite of these is the Kb, because it's so true. My iPod currently holds 1.75gb of music on its 2gig drive, and yet it still tells me it's full. Alas. Sigh. /dramatics.

That one is actually Apple's doing, the iPod OS is stored on the internal drive. iTunes adjusts the maximum size to account for the loss of space.

I, for one, refused to use the term kibibyte. I was raised to believe that a kilobyte is 1024 bytes. Thats what I will continue to call it until the day I die. I refuse to change my lexicon just because the drive manufacturers want to pad their numbers.

und3rdark
Posts: 27
Joined: Mon Jan 14, 2008 7:25 pm UTC
Location: Canton, Michigan

### Re: "Kilobyte" Discussion

I'm sorry, but when people tell me they have x kibibytes of y, I have this almost irrepressible desire to say: "You store your y on dog food?"

PS: Firefox spell-check marks kibibyte as a mis-spelling.
"The impulse to discriminate is a feature of our brains. We look for patterns and make decisions based on them. Sometimes the patterns are illusions, and we come to irrational conclusions."

endolith
Posts: 229
Joined: Tue Jan 01, 2008 2:14 am UTC
Location: New York, NY
Contact:

### Re: "Kilobyte" Discussion

kriel wrote:But seriously, I see a lot of people in here voting for popularizing the idea of powers of 10 for computer spaces.

Here's the problem. EVERYTHING about a computer is based on powers of 2.

Internally. So what? The "atoms" of a computer are binary, but the things that are made out of those atoms - the ways we interact with computers - have nothing to do with binary.

A photo is stored on your hard drive as binary bits, in chunks of a power-of-two size based on a binary address space. But the amount of data contained in that file has no relation to powers of two. This image on my desktop is 325,273 bytes. Where's the connection to binary?

For a typical user, the only information they need about this file's size can be represented as "325 kB". The apt-get system in Linux, for instance, uses this convention when reporting package sizes. The Linux kernel uses this convention when reporting hard drive sizes. etc. etc. There's no need to use the new binary "KiB" units in most situations. Just use the standard SI units correctly.

If we're really going to revert back to 'True' metric, using powers of 10, knock EVERYTHING back down to bits and start using scientific notation.

First, there's no "reversion". We've been using metric all along. Dig around in old computer documents on bitsavers. Drum memories were measured in decimal, just like modern hard drives. 8-inch floppies were measured in decimal. Speed units like megahertz or megabit/second have always been measured in decimal. Kids who grew up with Commodore 64s or took their first computing courses in the days of 5.25" floppy drives think that the KB=1024 convention is the only way it's ever been done, but that's not actually true. Even in those days it was not used consistently. Just look at the "1.44 MB" floppy.

Second, if we convert everything to decimal, as it probably should be for consistency and usability, it will only be on the interface side of things. The internals will still be binary logic with binary addressing and binary counters, but the user won't have to worry about any of that, since it has nothing to do with the things he's doing on the computer. The only thing I can think of that I would ever use that should report in powers of two is a partition manager. GNOME partition manager does exactly this, using the new standard binary prefixes.
Last edited by endolith on Mon Mar 10, 2008 2:57 pm UTC, edited 2 times in total.

Quan
Posts: 71
Joined: Mon Aug 20, 2007 6:17 pm UTC
Location: Manchester, UK

### Re: "Kilobyte" Discussion

I've actually been doing a lot of reading in to this recently because I've been trying to convince Firefox to change their notation. Under the European Harmonisation Document and the new up and coming ISO 80000, if we're to be technical about things:

KB = Kelvin-Byte
kB = kilobyte (10^3 bytes)
KiB = kibibyte (2^10 bytes)
B = Byte = n bits acted on as a group
o = Octet= 8 bit byte

I don't believe it's pointless and pedantic, it causes real confusion among users when a hard drive is right advertised as 500 GB and then Microsoft show the space in GiB.
I'm so geeky I got really excited about making this signature in SVG until it occurred to me HTML would obviously be turned off ¬_¬'

daranz
Posts: 5
Joined: Wed Apr 04, 2007 6:19 pm UTC
Location: Variable

### Re: "Kilobyte" Discussion

I say we all start using kiloquads. They're nice, because you can use arbitrary numbers and pretend they're true.
//TODO: Sig

pinkgothic
Posts: 140
Joined: Sat Dec 29, 2007 5:47 pm UTC
Location: Germany
Contact:

### Re: Data rates

DragonHawk wrote:[...] Bit rates apply to serial communications; byte rates vary with usage. [...]

I'm torn between worshipping you for the truth in your words and clobbering you over the head so you never ruin any future opportunity of mine to glow with knowledge and feel important.

I think the worshipping option is less hassle.

As for a solution to base-ten and base-two: KB2 vs. KB10. Simple and not confusing, though if you want to churn out the pedants then KB2 by analogy would have to be 8 bytes, not 1024 bytes, and then we're back at confusing. (Cookies to the person who figures out why and doesn't trip over their tongue explaining it.)

Or we start labelling boxes with: "Caution: May use binary phrases in base ten."

Or, once we're sick of it all: "Caution: May contain bobcat."
Stuff: own scribblings; favourite mewsic: luxuria & @440; underrated game

Quan
Posts: 71
Joined: Mon Aug 20, 2007 6:17 pm UTC
Location: Manchester, UK

### Re: Data rates

pinkgothic wrote:As for a solution to base-ten and base-two: KB2 vs. KB10. Simple and not confusing, though if you want to churn out the pedants then KB2 by analogy would have to be 8 bytes, not 1024 bytes, and then we're back at confusing. (Cookies to the person who figures out why and doesn't trip over their tongue explaining it.)

I think it's less worthwhile to come up with an entirely new consistent notation involving numbers and sub or supper notation and just at that point measure everything exactly in terms of bits or nats.
I'm so geeky I got really excited about making this signature in SVG until it occurred to me HTML would obviously be turned off ¬_¬'

hobbesmaster
Posts: 239
Joined: Sat Sep 15, 2007 10:11 pm UTC
Location: Lexington, Ky

### Re: "Kilobyte" Discussion

dr7 wrote:Look, people. I don't care about what "standard" notation for kilo/mega/etc are. This is computer world!

See, as an engineer I'd rather follow IEEE (1541) and IEC (60027) standards than some convention. (incidentally, JEDEC confuses things by having their own definition of KB, MB and GB for use in memory. note that KB != kilobyte or kibibyte or anything else)

Posts: 240
Joined: Mon Dec 11, 2006 4:22 pm UTC
Contact:

### Re: Data rates

Ah, I love strips like this. Every line is a gem.

I think I like the KBa the most. It leaves room for a parity check.

Re: Powers of 2 or powers of 10, it really depends on what perspective you come from. If you're doing FFTs, I'm pretty sure you care way more about powers of 2.

(Personally, I consider kilometers to be 1024 meters )

'; DROP DATABASE;-- wrote:Also, memory chips. 8 megabit? Why the hell do I care about the size in bits? Just say 1 megabyte.

Memory chips are a strange beast; they can have varying word sizes (8, 16) and/or parity (9, 18). I prefer to memory to be defined in terms of two quantities, data bus width and address bus width. One megabyte isn't as useful as knowing it's 512 rows of 16 bit words.

You can say that again! (= double word)

JamesCFraser
Posts: 32
Joined: Wed Nov 14, 2007 9:47 pm UTC

### Re: "Kilobyte" Discussion

Linux0s wrote:The FPU Intel kb is funny because it's true.

Blatantly not.

Assuming just a 16-bit normalised signed floating point number with 8-bits each for the exponent and the mantissa you can represent 1024 exactly.

i.e:

0100 0000 0000 1011

= 0.12 * 10210112
= 10,000,000,0002
= 102410

zenten
Posts: 3799
Joined: Fri Jun 22, 2007 7:42 am UTC

### Re: "Kilobyte" Discussion

hobbesmaster wrote:
dr7 wrote:Look, people. I don't care about what "standard" notation for kilo/mega/etc are. This is computer world!

See, as an engineer I'd rather follow IEEE (1541) and IEC (60027) standards than some convention. (incidentally, JEDEC confuses things by having their own definition of KB, MB and GB for use in memory. note that KB != kilobyte or kibibyte or anything else)

You engineers have been messing up computers since day one.

ubaphygg
Posts: 3
Joined: Sun Feb 03, 2008 4:05 pm UTC

### Re: "Kilobyte" Discussion

I don't understand why this stirred such a reaction with so many.
I'm sure Randall was just KiBytzing.

Ok I'll go now.

zahlman
Posts: 638
Joined: Wed Jan 30, 2008 5:15 pm UTC

### Re: Data rates

You can say that again! (= double word)

To make a long story short, that isn't funny.
Last edited by zahlman on Mon Mar 10, 2008 5:22 pm UTC, edited 1 time in total.
Belial wrote:I once had a series of undocumented and nonstandardized subjective experiences that indicated that anecdotal data is biased and unreliable.

Zake
Posts: 77
Joined: Thu Aug 16, 2007 2:25 am UTC

### Re: "Kilobyte" Discussion

Y'know, xkcd exemplifies a strange sort of sophisticated, yet so humble aesthetics. And one element of these aesthetics is that, while all of the text in this table is handwritten, the lines of the table itself are Powerpoint-table sharp. Thats ingenious. Xkcd would simply not be the same typed, or if Randall hand-wrote the table lines.

Details matter.

pinkgothic
Posts: 140
Joined: Sat Dec 29, 2007 5:47 pm UTC
Location: Germany
Contact:

### Re: Data rates

*grabs a bigger shovel*

zahlman wrote:To make a long story short, that isn't funny.

Bolded for truth.

Stuff: own scribblings; favourite mewsic: luxuria & @440; underrated game

Quan
Posts: 71
Joined: Mon Aug 20, 2007 6:17 pm UTC
Location: Manchester, UK

### Re: "Kilobyte" Discussion

hobbesmaster wrote:
dr7 wrote:Look, people. I don't care about what "standard" notation for kilo/mega/etc are. This is computer world!

See, as an engineer I'd rather follow IEEE (1541) and IEC (60027) standards than some convention. (incidentally, JEDEC confuses things by having their own definition of KB, MB and GB for use in memory. note that KB != kilobyte or kibibyte or anything else)

KB = Kelvin-Byte

The well known unit of heat information
I'm so geeky I got really excited about making this signature in SVG until it occurred to me HTML would obviously be turned off ¬_¬'

Posts: 240
Joined: Mon Dec 11, 2006 4:22 pm UTC
Contact:

### Re: "Kilobyte" Discussion

JamesCFraser wrote:
Linux0s wrote:The FPU Intel kb is funny because it's true.

Blatantly not.

Assuming just a 16-bit normalised signed floating point number with 8-bits each for the exponent and the mantissa you can represent 1024 exactly.

i.e:

0100 0000 0000 1011

= 0.12 * 10210112
= 10,000,000,0002
= 102410

1) You didn't bias your exponent.
2) After normalizing, the mantissa would be 0, because a pure power of 2 can be represented solely by the exponent's value. The normalizing procedure would trade the 0.12 in for a decremented exponent.

IEEE-754-1985 representation of the integer 1024 would have a 0 sign, 137 exponent, 0 mantissa. The binary representation would be 0 1000 1001 000 0000 0000 0000 0000 0000, and the hex representation 0x44800000.

pinkgothic wrote:
zahlman wrote:To make a long story short, that isn't funny.

Bolded for truth.

Is it still serious if it's a long, long story?

I'm ashamed to admit I didn't get it until you bolded it.

Quan wrote:KB = Kelvin-Byte

The well known unit of heat information

Is this unit used to measure the intensity of a flame war?

jerome_bc
Posts: 365
Joined: Wed Dec 12, 2007 1:12 pm UTC
Location: above
Contact:

### Re: Data rates

pinkgothic wrote:As for a solution to base-ten and base-two: KB2 vs. KB10. Simple and not confusing, though if you want to churn out the pedants then KB2 by analogy would have to be 8 bytes, not 1024 bytes, and then we're back at confusing. (Cookies to the person who figures out why and doesn't trip over their tongue explaining it.)

If KB10 = 103 = 1000
Then logically KB2 = 23 = 8

Kibi or kilo, I don't really care either way. I agree that using kibi for binary and kilo for decimal would be clearer, but I'm used to kilo being used both ways so I'm not confused by it anymore.

Quan
Posts: 71
Joined: Mon Aug 20, 2007 6:17 pm UTC
Location: Manchester, UK

### Re: "Kilobyte" Discussion

Quan wrote:KB = Kelvin-Byte

The well known unit of heat information

Is this unit used to measure the intensity of a flame war?

Strangely makes sense, given a constant co-efficient of heat information for a thread, as heat goes up due to more flaming, information goes down.
I'm so geeky I got really excited about making this signature in SVG until it occurred to me HTML would obviously be turned off ¬_¬'

pinkgothic
Posts: 140
Joined: Sat Dec 29, 2007 5:47 pm UTC
Location: Germany
Contact:

### Re: Data rates

jerome_bc wrote:Do I get a cookie?

Certainly. Do you take poisoned chocolate chip flavour?

Hey, I said nothing about it being edible!

* flees *
Stuff: own scribblings; favourite mewsic: luxuria & @440; underrated game

swirl
Posts: 1
Joined: Mon Mar 10, 2008 7:50 pm UTC

### IMO...

Here are a few points worth considering:

* kilo=1024 will always be confusing since the word originally and in all other fields was/is defined as kilo=1000 (part of me even screams: get your dirty 2-base fingers off of that SI defined prefix!)
* the difference between base-2 and base-10 prefixes grows (2, 4, 7, 9, 12 percent for kilo, mega, giga, tera and peta respectively)
* many things on computers don't make (more) sense in 2-base than in 10-base (filesystem/hard drive sizes: no, file sizes: no, amount you downloaded of the internet the last month: not that either, no)

PtS
Posts: 99
Joined: Fri Apr 13, 2007 9:17 am UTC
Location: Finland (the swedish part)
Contact:

### Re: "Kilobyte" Discussion

Simply put, best comic for a long time.
My life is 50% sarcasm and 50% sick humour.

-Me (said with 50% sarcasm)

ohki
Posts: 187
Joined: Wed Aug 23, 2006 9:27 am UTC
Location: San Luis Obispo, California
Contact:

### Re: "Kilobyte" Discussion

JamesCFraser wrote:
Linux0s wrote:The FPU Intel kb is funny because it's true.

Blatantly not.

Assuming just a 16-bit normalised signed floating point number with 8-bits each for the exponent and the mantissa you can represent 1024 exactly.

i.e:

0100 0000 0000 1011

= 0.12 * 10210112
= 10,000,000,0002
= 102410

http://en.wikipedia.org/wiki/Pentium_FDIV_bug
But it raining and me peeing on your foot are NOT mutually exclusive.
"Isn't arrogance measured in nano-Dijkstra's?"- Alan Kay

OliverKlozoff
Posts: 3
Joined: Mon Mar 10, 2008 9:28 pm UTC

### Re: "Kilobyte" Discussion

Foone wrote:I was bored and sick so I implemented the comic.
For example, when run on a music video by Herman's Hermits, I get the following file sizes:

Code: Select all

foone@mobile:~/Desktop\$ python howbig.py Movies/henry.flv Movies/henry.flv  4063 kB  4015 KB  (3968.69348078-3.87567722733j) KiB  3965.06702992 kb  4475 Kb  3527 KBa

For "Kb", does your script stat() the file and use the year portion of its mtime to determine the value for the units? e.g. 1024 - 4 * (year - 1980)