0394: "Kilobyte"

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

Moderators: Moderators General, Prelates, Magistrates

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

Re: "Kilobyte" Discussion

Postby OliverKlozoff » Mon Mar 10, 2008 9:52 pm UTC

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.


Kilobytes and Megabytes weren't ambiguous until some bastards started intentionally misinterpreting 'megabyte' as 10**6 instead of 2**20.

I have yet to see any software, *anywhere*, that claims that 1KB == 1000 bytes. It wasn't until HD mfgs started making very large drives that they started mucking up the term 'megabyte' to mean 10**6 instead of 2**20 bytes. I don't know who did it first, but I"m pretty sure it went like this - - Company X started futzing with the numbers to make their drives seem 2% larger than the others, and *every other company* followed suit to match. Eventually they managed to somehow convince themselves that misleading the public was the right thing to do.

Do you doubt me? Let's take a look at the original 10MB hard drive IBM produced for their IBM PC:

http://stason.org/TULARC/pc/hard-drives ... ST412.html

4 heads * 306 cylinders (tracks) * 17 sectors/track * 512 bytes/sector == 10653696 bytes, which rounds down to 10 * 2**20 bytes.

tgape
Posts: 57
Joined: Wed Oct 10, 2007 5:18 pm UTC

Re: "Kilobyte" Discussion

Postby tgape » Mon Mar 10, 2008 10:05 pm UTC

phlip wrote: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.


Also note that there have been problems with at least some attempts at making, for example, 384MB memory sticks. Issues I've noted included some motherboards only recognizing the last 1/3 of the memory if the stick was in the "last" memory slot, some motherboards would seize up if you stuck two such memory sticks in it, some motherboards just wouldn't recognize the last 1/3 of the memory, and one motherboard I had would decrease the memory bandwidth to 50% of ideal if one or more of these chips was present (that motherboard wanted 'matched' memory in every pair of slots. If any slot was filled with a different type of memory than its pair, or if it had a stick which did not contain an integer 2^n bits, it accessed memory in a degraded mode.)

I suspect that other motherboards could easily have other issues when presented with such memory, but fortunately, my time in hardware support ended, and I encountered no others.

phlip wrote: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.


Except when it wasn't. I'm pretty sure I remember seeing adverts for two different processors which *happened* to have just over an integer number of 1024*1024 Hz, and the manufacturer tried a small run of commercials pointing out their 1024 kilo compliance. They, um, didn't do so well. Apparently, 5% isn't a significant win - especially when everyone's rounding to the nearest tenth anyway.

phlip wrote: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.


Even when the defacto standard is no standard. If memory serves me correctly, hard drives were actually inconsistently marketed on this until somewhere between 40MB and 250MB. Note: I also missed the first hard drives; it's possible they were consistently labeled with the 2**20 MB rather than 10**6. (I was alive for the first hard drives, I just wasn't paying attention yet.)

Also: Very yes to the "thank god baud is gone" motion.


It seems most people would like to forget all about stop bits and all of the moments of silence by which Ethernet lives, as I don't believe anyone's mentioned them before me.


So, um, anyone up for changing our numerical base to better work with computers? I'm fine with octal or hex - although I think if we're going to seriously change to using hex, we need six new symbols, rather than just continuing to overload A through F...

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

Re: "Kilobyte" Discussion

Postby Quan » Mon Mar 10, 2008 10:09 pm UTC

OliverKlozoff wrote:
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.


Kilobytes and Megabytes weren't ambiguous until some bastards started intentionally misinterpreting 'megabyte' as 10**6 instead of 2**20.

I have yet to see any software, *anywhere*, that claims that 1KB == 1000 bytes. It wasn't until HD mfgs started making very large drives that they started mucking up the term 'megabyte' to mean 10**6 instead of 2**20 bytes. I don't know who did it first, but I"m pretty sure it went like this - - Company X started futzing with the numbers to make their drives seem 2% larger than the others, and *every other company* followed suit to match. Eventually they managed to somehow convince themselves that misleading the public was the right thing to do.


Does it really matter? Standardisation is a useful thing, having the prefix mega- always meaning the same thing is good. Not ending up with stupid situations like the 1.44 MB floppy disk (1.44 * 10^3 * 2^10) is good.

And I don't think I've ever seen any software equate 1 KB to 1000 bytes, but this is useful in 2 ways:

1) KB is not the SI for kiloByte

2) Computer users are used to things being measured in powers of 2 on their computer, so KiB is the natural progression.

Eventually the difference between the SI and IEC prefix will approach any arbitrarily large % clarity is good.
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 ¬_¬'

Aniviller
Posts: 23
Joined: Fri Oct 19, 2007 8:14 pm UTC

Re: "Kilobyte" Discussion

Postby Aniviller » Mon Mar 10, 2008 10:21 pm UTC

Totally sad and true comic, I get mad every time somebody turns up with an idea of 1kB=1000B
I say that unit Byte itself IMPLIES the meaning of k=1024. I can't be anything else. SI shouldn't mess with digital units, they just don't get it. And I'll walk on the moon before I use funny nonsense like KiB. It really looks like imaginary unit.

endolith
Posts: 227
Joined: Tue Jan 01, 2008 2:14 am UTC
Location: New York, NY
Contact:

Re: "Kilobyte" Discussion

Postby endolith » Mon Mar 10, 2008 10:25 pm UTC

OliverKlozoff wrote:Kilobytes and Megabytes weren't ambiguous until some bastards started intentionally misinterpreting 'megabyte' as 10**6 instead of 2**20.

Oh, conspiracy theories are such fun.

I have yet to see any software, *anywhere*, that claims that 1KB == 1000 bytes.

How long have you been using computers?

Do you doubt me? Let's take a look at the original 10MB hard drive IBM produced for their IBM PC:

What are you trying to prove with this? That hard drives have always been measured in decimal quantities since the beginning of time, since it's the most logical way to measure a quantity that has no inherent connection to binary multiples? Let's look at some more old computer hardware:

"This operation check detects errors in programming that cause invalid addresses. Examples: 40,000-and-above on a 40K core array; 20,000-and-above on a 20K core array. On a 10K core array, invalid addresses are detected by the address-bus validity check." http://www.bitsavers.org/pdf/ibm/14xx/1 ... _Oct61.pdf

"4k character storage", "4k IBM 1401" = 4,000 character memory http://portal.acm.org/citation.cfm?id=369093

Memorex 650 8-inch "1.5 megabit" floppy: = 1,500,000 bits

"1401 features supported are advanced programming, sense switches, tapes, multiply, divide, 16K core" ... "16,000 bytes"

Yep. K has definitely meant 1024 since the beginning of time, and was only recently bastardized to mean 1000 for marketing purposes. :roll:

User avatar
DeadCatX2
Posts: 240
Joined: Mon Dec 11, 2006 4:22 pm UTC
Contact:

Re: "Kilobyte" Discussion

Postby DeadCatX2 » Mon Mar 10, 2008 10:31 pm UTC

tgape wrote:
Also: Very yes to the "thank god baud is gone" motion.

It seems most people would like to forget all about stop bits and all of the moments of silence by which Ethernet lives, as I don't believe anyone's mentioned them before me.

I thought the baud went the way of the dodo due to being defined as the number of symbols per second. This can be extremely disingenuous; stop symbols count towards the baud rate, but not the bit rate. (bit rate slightly less than baud rate) More importantly, technologies like quadrature amplitude modulation can result in a symbol that represents more than than a single bit (bit rate much higher than baud rate). I vaguely recall Manchester encoding somehow making the concept of a baud awkward (the symbol rate for Manchester encoding schemes is not constant, or something?)

"Apples to oranges" == "bits to bauds"

User avatar
Bosonator
Posts: 21
Joined: Mon Jan 21, 2008 3:50 pm UTC

Re: "Kilobyte" Discussion

Postby Bosonator » Mon Mar 10, 2008 11:24 pm UTC

Who is "Kelly-Bootle"? Is that a real thing? It sounds like a joke, but sometimes reality is just weird enough that it can be funny just by "telling it as it is". :)
God I can handle... it's His fan club that scares me.

willamet
Posts: 1
Joined: Mon Mar 10, 2008 11:18 pm UTC

Re: "Kilobyte" Discussion

Postby willamet » Mon Mar 10, 2008 11:25 pm UTC

I propose a new standard that has more usefulness, I call it the milobyte. Specifically, its the number of bytes require to represent an orange tabby cat on the web. It's a great standard in that it moves with the times. In the old days, it was a paragraph of text, now its a video.

Extra bonus, Milo passed a way a while back and I had him cremated so he can also double as a weight and volume standard.

Kalos
Posts: 172
Joined: Thu Jan 31, 2008 6:45 pm UTC

Re: "Kilobyte" Discussion

Postby Kalos » Tue Mar 11, 2008 12:26 am UTC

Ari wrote:I believe you're specifically referring to a mean. It can't be just any average, as a median or a mode are both impossible to take from just two values, and in fact a mode is impossible to take from any set that involves purely unique values. ;)

Well yes, since the other two are both impossible to take from merely two values... I'm clearly referring to the only remaining type.

Jvr_Rz
Posts: 4
Joined: Sun Feb 03, 2008 11:39 pm UTC

Re: "Kilobyte" Discussion

Postby Jvr_Rz » Tue Mar 11, 2008 12:56 am UTC

As a matter of fact, lately one of the major ISPs here in Mexico City ("Cablevision") was really messing up... offering you a cheap transfer rate of "1.5 megabytes per second" on TV ads, when they actually meant 1.5 megabits ps (MB /Mb).

I wanted to demand for such service at the price announced (or sue them or something), but I was too lazy (sues in Mexico don't work as nicely as I understand they do in the US). They recently changed their ad and say it correctly.

I hate it when ads make technical or spelling mistakes.

Fnordstrum
Posts: 6
Joined: Wed May 02, 2007 3:16 pm UTC
Location: Connecticut

Re: "Kilobyte" Discussion

Postby Fnordstrum » Tue Mar 11, 2008 1:19 am UTC

dr7 wrote: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?

:mrgreen:



I totally agree... I grew up without that silly sounding "kibi mibi" etc crap, it was understood that it was 1024 and not 1000... The only reason I can think of for people wanting to change it is because hard drive companies realized at some point that if they used 1000 bytes in a kilobyte instead of 1024, then they could advertise a bigger drive... And back in the day it wasn't quite so huge a difference as it is now, where you actually have 30 or 50 gigs less than the size it's "supposed" to be. (For decent sized drives)


Instead of coming up with stupid new terminology, we should make the hard drive companies use the correct units, seriously.

Fnord.

¤¤¤¤¤¤¤
Power = Work / Time
Knowledge is Power and Time is Money, so
Knowledge = Work / Money, and Money = Work / Knowledge
Therefore, as Knowledge approaches zero, Money approaches infinity (regardless of work done), and vice versa.

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

Re: "Kilobyte" Discussion

Postby Quan » Tue Mar 11, 2008 1:29 am UTC

Fnordstrum wrote:I totally agree... I grew up without that silly sounding "kibi mibi" etc crap, it was understood that it was 1024 and not 1000... The only reason I can think of for people wanting to change it is because hard drive companies realized at some point that if they used 1000 bytes in a kilobyte instead of 1024, then they could advertise a bigger drive... And back in the day it wasn't quite so huge a difference as it is now, where you actually have 30 or 50 gigs less than the size it's "supposed" to be. (For decent sized drives)


Instead of coming up with stupid new terminology, we should make the hard drive companies use the correct units, seriously.

Fnord.

¤¤¤¤¤¤¤


But it's actually not been the "correct" units for quite some time. And the ambiguity isn't a new thing, the 1.44 MB Floppy Disk is not ~1.44 * 2^20.

kilo, Mega, Giga etc.. has always had a constant meaning in science and using powers of 10 is occasionally useful in computing, so having clear definitions does make sense.

Things change, society progresses, no need to be a stick in the mud.
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 ¬_¬'

User avatar
VectorZero
Posts: 471
Joined: Fri Nov 02, 2007 7:22 am UTC
Location: Kensington

Re: "Kilobyte" Discussion

Postby VectorZero » Tue Mar 11, 2008 2:01 am UTC

Kalos wrote:
Ari wrote:I believe you're specifically referring to a mean. It can't be just any average, as a median or a mode are both impossible to take from just two values, and in fact a mode is impossible to take from any set that involves purely unique values. ;)

Well yes, since the other two are both impossible to take from merely two values... I'm clearly referring to the only remaining type.


Would you please explain why it is impossible to take the median from merely two values?
Van wrote:Fireballs don't lie.

Fnordstrum
Posts: 6
Joined: Wed May 02, 2007 3:16 pm UTC
Location: Connecticut

Re: "Kilobyte" Discussion

Postby Fnordstrum » Tue Mar 11, 2008 3:57 am UTC

Quan wrote:But it's actually not been the "correct" units for quite some time. And the ambiguity isn't a new thing, the 1.44 MB Floppy Disk is not ~1.44 * 2^20.

kilo, Mega, Giga etc.. has always had a constant meaning in science and using powers of 10 is occasionally useful in computing, so having clear definitions does make sense.

Things change, society progresses, no need to be a stick in the mud.


Eh, even if the prefixes as used for computer-related things aren't the same as how the prefixes are used for units in general, it's just how it's done, I wouldn't say it isn't the "correct" units.

I can see why some people might want a consistent definition of the prefixes across everything, but it's generally understood that when you're talking about bytes, unless you're a hard drive company, it's done in powers of 2 and not powers of 10, so I don't really see where there would be much confusion (As long as you understand the whole Hard Drives having a bigger number than actual thanks to good old marketing)


Plus, I personally like a bit of complexity to things, it's just more fun for a kilobyte to be 1024 bytes than to be a boring 1000 bytes....

So fnord.



Also, as far as averages go, I was always under the impression that if someone just says "average," they're talking about the mean, unless otherwise specified... Generally any of the other types of averages the 'average' (ha-ha) person wouldn't even be thinking about (if they even know about them), so it's assumed the word average by itself refers to the mean...

¤¤¤¤¤¤¤
Power = Work / Time

Knowledge is Power and Time is Money, so

Knowledge = Work / Money, and Money = Work / Knowledge

Therefore, as Knowledge approaches zero, Money approaches infinity (regardless of work done), and vice versa.

User avatar
phlip
Restorer of Worlds
Posts: 7543
Joined: Sat Sep 23, 2006 3:56 am UTC
Location: Australia
Contact:

Re: "Kilobyte" Discussion

Postby phlip » Tue Mar 11, 2008 4:53 am UTC

Fnordstrum wrote:Also, as far as averages go, I was always under the impression that if someone just says "average," they're talking about the mean, unless otherwise specified... Generally any of the other types of averages the 'average' (ha-ha) person wouldn't even be thinking about (if they even know about them), so it's assumed the word average by itself refers to the mean...

You mention the counterexample in your own explanation... whenever one refers to the "average person", they're generally referring to the mode, not the mean.

And whenever someone uses the joke "50% of <whatever> are below average", I like to pretened they're using the median (though, more likely, they're using the mean and are just probably wrong).

Code: Select all

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

synapse
Posts: 2
Joined: Tue Mar 11, 2008 9:52 am UTC

Re: "Kilobyte" Discussion

Postby synapse » Tue Mar 11, 2008 10:04 am UTC

Sorry, but Kba is taken. The geneticists got there first. We measure nucleotides in Kilobases, KBas, and megabases, MBas. Some people I've worked with in the lab write it down as Kb and Mb, and that bothers me.

User avatar
Foone
Posts: 17
Joined: Tue Nov 21, 2006 11:14 pm UTC
Location: Asheville, NC
Contact:

Re: "Kilobyte" Discussion

Postby Foone » Tue Mar 11, 2008 11:10 am UTC

OliverKlozoff wrote:
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)


No, it uses the current date. Doing it off the mtime doesn't seem to make much sense, that'd mean older files use bigger Kbs.

endolith
Posts: 227
Joined: Tue Jan 01, 2008 2:14 am UTC
Location: New York, NY
Contact:

Re: "Kilobyte" Discussion

Postby endolith » Tue Mar 11, 2008 12:05 pm UTC

Fnordstrum wrote:The only reason I can think of for people wanting to change it is because hard drive companies realized at some point that if they used 1000 bytes in a kilobyte instead of 1024, then they could advertise a bigger drive... And back in the day it wasn't quite so huge a difference as it is now, where you actually have 30 or 50 gigs less than the size it's "supposed" to be. (For decent sized drives)

Instead of coming up with stupid new terminology, we should make the hard drive companies use the correct units, seriously.


Yet another conspiracy theory criticism of hard drive companies without a shred of evidence to back it up.

The Finn
Posts: 23
Joined: Thu Mar 06, 2008 4:47 pm UTC

Re: "Kilobyte" Discussion

Postby The Finn » Tue Mar 11, 2008 3:30 pm UTC

endolith wrote:
Fnordstrum wrote:The only reason I can think of for people wanting to change it is because hard drive companies realized at some point that if they used 1000 bytes in a kilobyte instead of 1024, then they could advertise a bigger drive... And back in the day it wasn't quite so huge a difference as it is now, where you actually have 30 or 50 gigs less than the size it's "supposed" to be. (For decent sized drives)

Instead of coming up with stupid new terminology, we should make the hard drive companies use the correct units, seriously.


Yet another conspiracy theory criticism of hard drive companies without a shred of evidence to back it up.


I've been amassing disk images, /proving/ that the hard drive companies have been shipping disks with digital copies of Tibetan prayer wheels on them, in an attempt to generate karma to overthrow the government of the People's Republic. Now all I have to do is^G^G^G^G^G^G^G^G^H

this post has been censored by the Department of Homeland Security. Nothing (0 x 10**6) to see here. Move along. Remember, Citizens, your Boot Sectors are encrypted to protect you from int13h viruses! Keep Norton running at all times!

Soft Hyphen
Posts: 18
Joined: Tue Mar 11, 2008 4:10 pm UTC

Re: Intel jokes

Postby Soft Hyphen » Tue Mar 11, 2008 4:33 pm UTC

DragonHawk wrote:
netsplit wrote:Why can't networking equipment tell me the speed in bytes?

Because throughput in bytes depends entirely on the framing and protocols you're using, while bitrate is constant and unambiguous.

If you just want a number that makes you feel better, divide the bitrate by eight. Or seven, might as well make it go faster than possible, since you don't care about correctness.

I finally registered just to argue this point.

You are making the assumption that raw throughput can only be measured in bits, and bytes can only be used to measure goodput.

There is no reason for this kind of limitation.

1 megabit per second raw transfer rate = 128 kilobytes per second raw transfer rate
128 kilobytes per second of user data = 1 megabit per second of user data

There is no reason why networking equipment capability can't be quoted, unambiguously, in bytes per second.

If a network card tells me it can do 128 megabytes per second, there's no confusion. Yes, I won't be able to transfer 128 megabytes of "real" data every second because of protocol overhead, but so what? I can't transfer 100 "real" megabits either. It's less confusing for me to think "128MBps minus overhead cost" than to do mental arithmetic every damn time to convert from bits to bytes, and then factor in overhead cost anyway.

And 1 byte = 8 bits. It has been this way for some time now, get used to it. The metre used to be defined according to the circumference of the Earth, now it's based on the speed of light. No one argues with that. Why the argument that a "byte isn't necessarily 8 bits". Of course it is. That's its modern definition.

User avatar
fuzzball45
Posts: 46
Joined: Fri Sep 21, 2007 3:28 pm UTC
Location: The World, The Universe

Re: "Kilobyte" Discussion

Postby fuzzball45 » Tue Mar 11, 2008 5:03 pm UTC

endolith wrote:
OliverKlozoff wrote:Kilobytes and Megabytes weren't ambiguous until some bastards started intentionally misinterpreting 'megabyte' as 10**6 instead of 2**20.

Oh, conspiracy theories are such fun.


Conspiracy Theories
dmoney wrote:
russianspy1234 wrote:What does that make her to me?


Fair game.

killerstar
Posts: 24
Joined: Wed Apr 11, 2007 8:59 am UTC

Re: "Kilobyte" Discussion

Postby killerstar » Tue Mar 11, 2008 6:57 pm UTC

Instead of changing cumputer meassures to fit conventional ones, we should do exactly the oposite. 1 km = 1024 m, 1 m = 1024 cm and so on.

The world would be a much better place.

User avatar
Feirgon
Posts: 16
Joined: Fri Jan 18, 2008 3:33 pm UTC
Location: New York

Re: "Kilobyte" Discussion

Postby Feirgon » Tue Mar 11, 2008 7:14 pm UTC

<rant>
First of all, no matter the size of a statistical data set you can take the mean, median, and mode of said data. If all of your data is unique, you have a mulitmodel set (what? more than one mode for a data set OH NOES!). This rule applies to finding the median of even sets of numbers, both of the middle values are considered the median (you can take the mean of these two numbers, though).

Second, I hated statistics.

Third, I am a computer programmer. I have been around computers since grade school. I currently have a degree in Computer Science. I currently work as a Software Engineer.

WHEN DID 1000 = 1024???

This has always irked me. I'm not saying I have been actively fighting against this brutality of the SI system, I just have never truely accepted it. I am glad new standards are coming out, whether or not the current people out there will accept it, is moot. It will be taught to younger programmers and your ignorant disregard for "the meaning of words and numbers" will go the way of the dinasaurs. (Unless George Bush convinces us that 2 = 3 ...). Hmmm, I like that.

SAYING YOU THINK KILOBYTE MEANS 1024 BYTES IS LIKE SUPPORTING BUSH!!!

Wow, ignorant, ignitable statements are easy to come up with...
</rant>

Munroe, you have outdone yourself once again.
Image
Be the Ultimate Ninja! Play Billy Vs. SNAKEMAN today!

User avatar
DeadCatX2
Posts: 240
Joined: Mon Dec 11, 2006 4:22 pm UTC
Contact:

Re: "Kilobyte" Discussion

Postby DeadCatX2 » Tue Mar 11, 2008 7:35 pm UTC

Feirgon wrote:<rant> [..] WHEN DID 1000 = 1024???

This has always irked me. I'm not saying I have been actively fighting against this brutality of the SI system, I just have never truely accepted it. I am glad new standards are coming out, whether or not the current people out there will accept it, is moot. It will be taught to younger programmers and your ignorant disregard for "the meaning of words and numbers" will go the way of the dinasaurs.

<sarcasm>When did + = -? When doing the math, engineers always consider electrons to flow from positive to negative, but this is completely, factually incorrect.</sarcasm>

My prof would always say "it's a lie. But it's a good lie; it makes life happier for us. :) " We consider electricity to flow from + to - because it makes the math easier in that field. Similarly, considering k to be 210 has made the math easier for some fields of computers. I think ambiguity is a small price to pay for needing a calculator less often.

That said, I have done this before.

First Engineer: We'll need a 2 meg SRAM
Second Engineer: 2 meg, base 2 or base 10?
First Engineer: Base 2

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

Re: "Kilobyte" Discussion

Postby Quan » Tue Mar 11, 2008 8:13 pm UTC

DeadCatX2 wrote:My prof would always say "it's a lie. But it's a good lie; it makes life happier for us. :) " We consider electricity to flow from + to - because it makes the math easier in that field. Similarly, considering k to be 210 has made the math easier for some fields of computers. I think ambiguity is a small price to pay for needing a calculator less often.


Well that's o.k for engineers, not so much for mathematicians. Ambiguity is the absolute cardinal sin in mathematics
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 ¬_¬'

Dave54
Posts: 1
Joined: Tue Mar 11, 2008 7:40 pm UTC

Pentium F.P.U.

Postby Dave54 » Tue Mar 11, 2008 8:25 pm UTC

The Intel kilobyte, calculated on the Pentium F.P.U., has the wrong value in the comic. It should be 1023.9375228 (note the extra 2).

By far, the most famous Pentium F.P.U. miscalculation is 4195835/3145727. See http://en.wikipedia.org/wiki/Pentium_FDIV_bug#References.
4195835/3145727 equals 1.333820449136241002, but calculated on the Pentium F.P.U. it equals 1.333739068902037589.
The ratio of the two, 1.333739068902037589/1.333820449136241002, equals 0.9999389871146. Multiply that by 1024 and you get 1023.9375228.
Therefore, 1023.9375228 has the same error ratio as the Pentium F.P.U. However Munroe entered 1023.937528, which I assume is a typo.

To address the possibility of Munroe hitting a rounding error, I recalculated it rounding at every sensible decimal place:

Code: Select all

No Rounding:
1024*1.333739068902037589/1.333820449136241002 = 1023.9375228053534040874498553006

Rounding Down:
1024*1.33373906890203758/1.33382044913624100 = 1023.9375228053533987133195073731
1024*1.3337390689020375/1.3338204491362410 = 1023.9375228053533372957576061071
1024*1.333739068902037/1.333820449136241 = 1023.9375228053529534359957231923
1024*1.33373906890203/1.33382044913624 = 1023.9375228053483470720123448893
1024*1.3337390689020/1.3338204491362 = 1023.9375228053560223936186704466
1024*1.333739068902/1.333820449136 = 1023.9375228055095569302152003942
1024*1.33373906890/1.33382044913 = 1023.9375228085801539805936653332
1024*1.3337390689/1.3338204491 = 1023.9375228316103344707672618332
1024*1.333739068/1.333820449 = 1023.9375222174300313190055163105
1024*1.33373906/1.33382044 = 1023.9375229847279893236603871498
1024*1.3337390/1.3338204 = 1023.9375076284633223483461491507

Correct Rounding (rounding up 5 and above)
1024*1.33373906890203759/1.33382044913624100 = 1023.9375228053534063905147450317
1024*1.3337390689020376/1.3338204491362410 = 1023.9375228053534140677099826903
1024*1.333739068902038/1.333820449136241 = 1023.937522805353721155519489022
1024*1.33373906890204/1.33382044913624 = 1023.9375228053560242672500031918
1024*1.3337390689020/1.3338204491362 = 1023.9375228053560223936186704466
1024*1.333739068902/1.333820449136 = 1023.9375228055095569302152003942
1024*1.33373906890/1.33382044914 = 1023.9375228009034271507660175319
1024*1.3337390689/1.3338204491 = 1023.9375228316103344707672618332
1024*1.333739069/1.333820449 = 1023.9375229851495551632527115346
1024*1.33373907/1.33382045 = 1023.9375229851963958117451265638
1024*1.3337391/1.3338204 = 1023.9375844004185271120459696067

All the answers in a row:
1023.9375228053534040874498553006
1023.9375228053533987133195073731
1023.9375228053533372957576061071
1023.9375228053529534359957231923
1023.9375228053483470720123448893
1023.9375228053560223936186704466
1023.9375228055095569302152003942
1023.9375228085801539805936653332
1023.9375228316103344707672618332
1023.9375222174300313190055163105
1023.9375229847279893236603871498
1023.9375076284633223483461491507
1023.9375228053534063905147450317
1023.9375228053534140677099826903
1023.937522805353721155519489022
1023.9375228053560242672500031918
1023.9375228053560223936186704466
1023.9375228055095569302152003942
1023.9375228009034271507660175319
1023.9375228316103344707672618332
1023.9375229851495551632527115346
1023.9375229851963958117451265638
1023.9375844004185271120459696067


As you can see, it was not rounding error.

Now I can really relate to this comic:
Image

User avatar
Feirgon
Posts: 16
Joined: Fri Jan 18, 2008 3:33 pm UTC
Location: New York

Re: "Kilobyte" Discussion

Postby Feirgon » Tue Mar 11, 2008 8:44 pm UTC

Quan wrote:
DeadCatX2 wrote:My prof would always say "it's a lie. But it's a good lie; it makes life happier for us. :) " We consider electricity to flow from + to - because it makes the math easier in that field. Similarly, considering k to be 210 has made the math easier for some fields of computers. I think ambiguity is a small price to pay for needing a calculator less often.


Well that's o.k for engineers, not so much for mathematicians. Ambiguity is the absolute cardinal sin in mathematics


I have no problem with us thinking in base two, the problem was hijacking a word's already established meaning. I personally like the idea of coming up with new words to fit the meanings we want to convey...but that's just me (I am a word pioneer, give me a few drinks and you will hear word combinations you never thought of ;)
Image
Be the Ultimate Ninja! Play Billy Vs. SNAKEMAN today!

User avatar
errrr
Posts: 31
Joined: Tue Dec 04, 2007 9:18 pm UTC
Location: Swimming in a fish bowl, year after year

Re: "Kilobyte" Discussion

Postby errrr » Tue Mar 11, 2008 8:58 pm UTC

How could you miss the moronobyte? (defined as 1024*1000 byte) Wouldn't have fitted the kilo-theme though :)
"The Internet being what it is, absolutely anything might show up in the collage including -- quite possibly -- pornography, or even nudity." (xscreensaver/webcollage docs)

User avatar
Spoom
Posts: 109
Joined: Fri Feb 02, 2007 7:45 pm UTC

Re: "Kilobyte" Discussion

Postby Spoom » Tue Mar 11, 2008 9:48 pm UTC

All other things being equal, the simplest solution is ale.

User avatar
netsplit
Posts: 228
Joined: Fri Oct 12, 2007 3:34 am UTC
Location: Z̈̓̾̾̀̉ͬ̅ͯͪͬ̚͢͏̴̝̝̩̳͙͘a̷̧̰̥̯͎͍͍̻̮͔̭̰̗̺̩̐̀͛͐ͥ͐ͨ͆̀̉̐̌̅͑ͣ͟ͅľ̸̠̟̭̲̜̹̗͎͍̗̤͖̤͕̣̝̀̊̑ͯͬ͌̍ͬ͂ͤ͛̀̚g̍̅́ͮ̌̏ͣ́̀͏͍

Re: "Kilobyte" Discussion

Postby netsplit » Tue Mar 11, 2008 11:40 pm UTC

I don't see why SI definitions should apply to computer terms. Metric is nice for easy conversion, but I think of no circumstance where I'd have to convert meters to megabytes, or gigabytes to liters.

1,024 to the X works really nice for ram, maybe no advantages for hard drives, but I think using the same words having the same meanings works very nice. I mean going back to metric, it's like asking if it's "big liters" or "little liters"? I prefer liters just means liters. 1,024 powers work great for memory operations, but what technical gains are there for using powers of 1,000?
from da craddle to da grave, geek life 4 eva
better show hardcore respect ya'll

JayDee wrote:"What is the difference between erotic and kinky? Erotic is using a feather. Kinky is using the whole Dinosaur."

Puzzlemaker
Posts: 175
Joined: Thu Jul 26, 2007 4:16 pm UTC
Location: Maryland, Silver Spring, USA

Re: "Kilobyte" Discussion

Postby Puzzlemaker » Wed Mar 12, 2008 12:29 am UTC

WaitWaitWait.

Can someone please answer me this...

How can a kilobyte be referred to as 1024? Kilo means one thousand. A kilo-byte means one thousand bytes, not 1024. Unless it's talking about a kilobit, in which case it would be 125 Bytes. Or, 2^1000 bits, which would be a really big freaking number. Or, it might be talking about a thousand bytes in octal, which would be 1750. 1024 Bytes is 2000 in Octal, and 400 in Hex, so none of those match up, unless a kilobyte really means TwoKiloOctalBytes, or TKOB. Although, just KOB would be good, equaling 512 bytes. I don't know, I don't get it at all. Someone explain it to me?
RezardVareth wrote:Bad argumentation can be painful to answer, but responding with empty logic just perpetuates the cycle.

User avatar
phlip
Restorer of Worlds
Posts: 7543
Joined: Sat Sep 23, 2006 3:56 am UTC
Location: Australia
Contact:

Re: "Kilobyte" Discussion

Postby phlip » Wed Mar 12, 2008 12:48 am UTC

PuzzleMaker: 1024 is close enough to 1000 for most purposes (but not all purposes of course, which is where the problems come in)... and since it's a power of 2, it's a much nicer number for certain things in computers.

For instance: for assorted technical reasons, RAM will always be a power of 2 bits. It's cheaper that way, and it makes the motherboard chipsets easier to make (tgape pointed out above that many motherboards will fail if you put a non-power-of-2 RAM stick in). So if you get 1GB of ram, that's really 230 bytes.

This gets called 1GB because it's simpler, and usually close enough. Pedants (and I use the term with no pejorativity intended) complained that it was inaccurate and ambiguous to have the same terms for both binary and decimal prefixes, and introduced "kibi-", "mebi-", "gibi-", etc, to replace them. They haven't exactly taken off... but they are getting used a bit... I think they will eventually take over.

Code: Select all

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

endolith
Posts: 227
Joined: Tue Jan 01, 2008 2:14 am UTC
Location: New York, NY
Contact:

Re: "Kilobyte" Discussion

Postby endolith » Wed Mar 12, 2008 12:54 am UTC

DeadCatX2 wrote:My prof would always say "it's a lie. But it's a good lie; it makes life happier for us. :) " We consider electricity to flow from + to − because it makes the math easier in that field.


No. The math would be the same regardless of the polarity. We use + and − arbitrarily, because it doesn't matter which way it actually goes.

Current in metals flows as electrons from − to +, but current in semiconductors flows from + to − in the form of electron holes, current in ice flows as protons from + to −, and current in electrolyte solutions and ionized air can flow in both directions at once. The type of particles that move, the speed at which they move, and so on are all variable depending on conductor cross-section, conductivity, etc. etc. etc. So instead of worrying about all that, we just count the amount of net charge that passes a given point per unit time and call it current.

It's not a lie; it's a convention.

Similarly, considering k to be 210 has made the math easier for some fields of computers. I think ambiguity is a small price to pay for needing a calculator less often.


No one has a problem with using base 2. They have a problem with using base 10 units to refer to base 2 numbers.

Best description I've read of the unit issue:
http://meta.ath0.com/2005/02/23/a-plea-for-sanity/
Last edited by endolith on Wed Mar 12, 2008 7:18 am UTC, edited 2 times in total.

Puzzlemaker
Posts: 175
Joined: Thu Jul 26, 2007 4:16 pm UTC
Location: Maryland, Silver Spring, USA

Re: "Kilobyte" Discussion

Postby Puzzlemaker » Wed Mar 12, 2008 1:49 am UTC

I typed up a large paragraph twice now, and erased it, then stared at the blank text box and started anew. So, instead, I shall just summarize my feelings in a few succinct words.

What the hell?
RezardVareth wrote:Bad argumentation can be painful to answer, but responding with empty logic just perpetuates the cycle.

scarletmanuka
Posts: 525
Joined: Wed Oct 17, 2007 4:29 am UTC
Location: Perth, Western Australia

Re: "Kilobyte" Discussion

Postby scarletmanuka » Wed Mar 12, 2008 7:10 am UTC

endolith wrote:Current in metals flows as electrons from + to -, but current in semiconductors flows from - to + in the form of electron holes, current in ice flows as protons from - to +, and current in electrolyte solutions and ionized air can flow in both directions at once.

The semantics of - and + appear to be reversed in the above post.

endolith
Posts: 227
Joined: Tue Jan 01, 2008 2:14 am UTC
Location: New York, NY
Contact:

Re: "Kilobyte" Discussion

Postby endolith » Wed Mar 12, 2008 7:15 am UTC

scarletmanuka wrote:The semantics of - and + appear to be reversed in the above post.


Hah. Fixed. (And now with real Unicode minus signs!)

User avatar
DeadCatX2
Posts: 240
Joined: Mon Dec 11, 2006 4:22 pm UTC
Contact:

Re: "Kilobyte" Discussion

Postby DeadCatX2 » Wed Mar 12, 2008 4:00 pm UTC

endolith wrote: [..] current in semiconductors flows from + to − in the form of electron holes, [..]

It's not a lie; it's a convention.

You know that a hole is just an abstraction to describe a physical phenomenon in terms that are easier to grasp, right? In other words, holes are a lie, too.

But they're a good lie! :lol:


Puzzlemaker wrote:I typed up a large paragraph twice now, and erased it, then stared at the blank text box and started anew. So, instead, I shall just summarize my feelings in a few succinct words.

What the hell?

Let's assume that you want to make a RAM chip. We'll say it's a 2kB SRAM with 8 bit words. That means you will have 8 data pins. What about your address bus? A 2 kB SRAM will require an address bus width of 11 bits. That means there will be 211 individually addressable bytes in the SRAM.

So, when you have an SRAM with 2,048 bytes, how would you like to refer to that? I was raised to consider that to be 2 kilobytes. Byte, by its very nature, implies to me that the corresponding units are base 2.

Aside from years of conditioning, I think the only reason I don't like kibibyte is because it sounds really, really stupid. ("mebibyte" - mebi it's a byte, mebi it's not?) I don't hate kiB or MiB as much, but in my head I still say "kilo" and "mega" when I read them.

endolith
Posts: 227
Joined: Tue Jan 01, 2008 2:14 am UTC
Location: New York, NY
Contact:

Re: "Kilobyte" Discussion

Postby endolith » Wed Mar 12, 2008 5:06 pm UTC

DeadCatX2 wrote:You know that a hole is just an abstraction to describe a physical phenomenon in terms that are easier to grasp, right? In other words, holes are a lie, too.

You know that electrons are actually waves, right? :)

So, when you have an SRAM with 2,048 bytes, how would you like to refer to that? I was raised to consider that to be 2 kilobytes.

It is 2 kilobytes. 2.048 kilobytes, to be precise. The confusion started when programmers got to 21×1024 and called it "21K" instead of "22k".

Aside from years of conditioning, I think the only reason I don't like kibibyte is because it sounds really, really stupid.

Because "bite" and "nibble" are not unusual sounding at all... Years of conditioning, indeed.

User avatar
DeadCatX2
Posts: 240
Joined: Mon Dec 11, 2006 4:22 pm UTC
Contact:

Re: "Kilobyte" Discussion

Postby DeadCatX2 » Wed Mar 12, 2008 5:45 pm UTC

endolith wrote:
DeadCatX2 wrote:You know that a hole is just an abstraction to describe a physical phenomenon in terms that are easier to grasp, right? In other words, holes are a lie, too.

You know that electrons are actually waves, right? :)

Yes, I do. My view is that it's okay to believe a lie if it makes life easier for you. This is entirely consistent with knowing electrons are more accurately modeled by the Wave Function, but preferring to use more convenient "lies" (= abstractions) because it makes life less difficult.

endolith wrote:
DeadCatX2 wrote:Aside from years of conditioning, I think the only reason I don't like kibibyte is because it sounds really, really stupid.

Because "bite" and "nibble" are not unusual sounding at all... Years of conditioning, indeed.

Byte sounds perfectly reasonable; it even sounds like another word that exists in language. I think the part that is annoying is the bibyte part. It just doesn't flow.

Ar-Pharazon
Posts: 16
Joined: Sun Dec 23, 2007 3:50 pm UTC

Re: "Kilobyte" Discussion

Postby Ar-Pharazon » Wed Mar 12, 2008 10:43 pm UTC

Hahah, this comic is hysterical! By the way, it seems like the comics after Gygax's death (no, I'm not implying anything) are sillier and much less serious than those before.


Return to “Individual XKCD Comic Threads”

Who is online

Users browsing this forum: hetas, Yahoo [Bot] and 39 guests