Simon & Garfunkel; the gigabyte dichotomy
Sep. 21st, 2003 12:20 pm![[personal profile]](https://www.dreamwidth.org/img/silk/identity/user.png){kind=small}
robotech_masterLast week I saw Simon & Garfunkel perform "The Boxer" live on the David Letterman show. Wow. The magic, they've still got it...at least when they can stand each other long enough to perform together. In a way, though, their reunion concert tour is a double-shame: first, that they're not doing any new material, rather just rehashing the same old stuff all over again, and second, that the concert ticket prices are astronomical. Is it any wonder that people are trading whatever they can get for free?
Last night, I got into a big political argument with a friend of mine. Probably one of my first real such arguments ever, as I tend to stay largely detached from politics. It wasn't actually an argument about politics, it was just one of those bordering-on-religious debates where both sides know they're right, and yet they have to try to convince the other of that, too. It's the most insidious kind of argument...even now I have this feeling that if I maybe just restate what I'm saying a certain way, then my friend would say, "Oh, that's what you mean, I totally agree," because it just seems so sensible. But I know that it doesn't work that way.
What touched it all off is partly a lawsuit that was filed recently, and partly a recent purchase of mine. The lawsuit was a class-action suit filed on behalf of computer hard-drive purchasers who were disappointed that when they stuck their hard drives in the computer, the computer told them that there wasn't as much space on them as they thought they were getting.
I read the article, thought, "Huh," but it wasn't quite brought home to me what it was really all about until I got a new hard drive for my Wallstreet laptop. The old one was flaking out, and I was constantly having to rebuild the directory just to get it to boot. Given that I could either spend $100 and get a 20 gig drive, or spend $30 or so more and get a bigger and faster one, I figured a 40 gig drive was worth the extra expenditure. I thought about how it would be nice to have those gigs...I could spend 8 on OS X, since it needs to be installed within the first 8 gigs of the drive on my type of computer, and then have 32 left over for mp3s or other miscellaneous stuff.
Only, when I got the drive and put it in the computer, it actually registered as having 37.26 gigabytes, not 40...which means my drive actually only has about 93% of the actual space on it that it claims to. The reason for this is that the hard drive manufacturers choose to use a different definition of "gigabyte" than the rest of us do.
You see, in computers, kilo means 1024. This is because computers think in binary, powers of 2. Thus, 210 = 1024...and that's close enough to 1000 that, needing a good shortcut, the computer hackers of the day named 210 bytes a kilobyte. Subsequently, 1024 kilobytes became a megabyte, and 1024 megabytes a gigabyte. This means that, by the definition commonly used in computer science, a gigabyte is 1,073,741,824 bytes. This is a very fundamental definition in the computer industry, and it's taught to even the most basic computer science student. But technically, it is also semantically incorrect.
This is because the prefixes kilo-, mega-, and giga- have specific meanings in the scientific community outside the computer field—they refer to 1,000, 1,000,000, and 1,000,000,000 respectively, and are commonly used to tack onto other units to specify 103, 106, or 109 of that particular unit: kilogram, kilometer, kilocalorie, kiloton, and so on. The difference between 1,000 and 1,024 really isn't all that great...but it increases geometrically as the powers increase...which means there's a difference between a gigabyte in the computer sense and a gigabyte in the standards sense of 73,741,824 bytes, or 70.33 megabytes, or about 6.87%.
So, why is there a 7% difference in the size that the drive claims to be on the package and the size it claims to be in the computer? Why, because the hard drive manufacturing industry chooses to use the 109 definition of "giga" instead of the 1,073,741,824 definition used by the entire rest of the industry! If you buy a 40 gigabyte drive, as I did, you're actually getting a drive that can hold 40,000,000,000 bytes, not a drive that can hold 40 true gigabytes.
70 megabytes per gigabyte may not seem like much—that's about 1/10 of a CD-ROM, 7/10 of a Zip-100 disk—but if you multiply it by 100, you begin to realize there's a pretty big difference there. And now that it's common for consumers to buy 150 or 200 gigabyte drives and get "short-changed" by 10 or 15 gigabytes, it's not too surprising they're suing. I'm a bit annoyed myself, and at myself, for not realizing it before I bought this 40 gig drive and lowering my expectations accordingly.
Anyway, the thrust of the argument my friend and I were having is that I feel the hard drive manufacturers are engaging in a deceptive practice, puffing up the size of their hard drives by using a different definition than everybody else in the industry in a manner not unlike candy bar manufacturers who put their bars in a package bigger than the bar itself. They're in the computer industry, there's no way they can avoid knowing how big the drive will claim to be in someone's computer. They should label their drives accordingly—consumers should have a reasonable expectation of getting a drive their computer agrees is 150 gigabytes when they buy a 150 gigabyte drive.
My friend, on the other hand, apparently feels that the drive manufacturers are using the correct definition of "kilo" and everybody else isn't—that using kilo to mean 1024 somehow dilutes the meaning of the standard and confuses people. He's not the only one to feel this way, of course; a standards group has come up with new names for kilobyte, megabyte, and gigabyte that they hope people will switch to using instead: kibibyte, mebibyte, and gibibyte. (Or, as I call them, kibble-byte, maybe-bite, and gibberish-byte.) My friend feels that in the court case, the hard drive manufacturers can bring expert witnesses forward and claim that they're using the correct definition of giga, and so it's okay.
I just don't think that's going to work too well, because the courts will look at the non-standard way in which kilo-, mega-, and giga- are used by the rest of the industry and decide that consumers should have a reasonable expectation that hardware manufacturers will follow the computer-technology (non-)standard when they're making computer parts.
Nor do I think it too likely that people will switch over to using kibi-, mebi-, and gibi-. Aside from the fact that they look and sound nonsensical, human nature is that when something erroneous or overcomplicated gets entrenched, it continues to be used through the ages even when people come up with something more correct or internally consistent. We're still using English measurements instead of the easier-to-convert metric system (though, granted, Europe's using metric more than we are). We're still on a 60 second minute, 60 minute hour, 24-hour (in two cycles of 12) day, 7 day week, 28/29/30/31-day month, 52-week or 12-month or 365.25-day year (and even that requires 'leap seconds' or skipped leap days every now and then) instead of coming up with something metric to replace it. We're still using the 19th-century QWERTY keyboard, which was designed to slow typists down so they wouldn't break their equipment, instead of Dvorak or one of the other systems designed to speed typists up. Kibi-, mebi-, and gibi- may be more accurate than kilo-, mega-, and giga-, but they came along about three decades of usage too late.
Unlike in French, which has an Academy to ride herd on the language, the meaning of words in English is created and changed by usage, not the other way around. So words and expressions that were originally incorrect become correct if enough people use them for long enough. It used to be a popular schoolyard expression to say, "Don't say 'ain't' 'cuz 'ain't' ain't a word!" But the expression itself is incorrect now, because "ain't" is enough of a word to have a lengthy definition in the Merriam-Webster Dictionary. Thus it is for kilobytes, megabytes, gigabytes, and so forth. But kibibytes and their ilk "ain't" in there yet.
Heck, it's not even the first time that "kilo" has been incorrectly used. Or, in this case, incorrectly not used. I learned back in my high school chemistry classes that there are actually two kinds of the unit of energy we call calories: calories and Calories with a capital C. The capital C Calories are actually 1,000 of the small-c calories, and are also called kilocalories. And guess which of these units is used to list the energy content of food? Kilocalories, or Calories. Except nobody but scientists calls them kilocalories, or bothers to capitalize when talking about "counting calories."
My friend feels that the courts aren't where this standards matter should be resolved...but I think they're the only place where the only part of it that can be resolved (getting the drive manufacturers to start labelling their products by the same definition everyone else uses) will be...and it had probably better be soon. The margin of error will only increase when they start selling terabyte (210 gigabytes) and petabyte (210 terabytes—the equivalent of twenty million four-drawer filing cabinets full of text) drives.
In the end, he and I simply had to agree to disagree, because there's no way we're going to be able to change each others' minds. (I still have this urge that pops up every so often to try it again, that surely he'll listen this time. But the wiser part of me says to just let it rest, and so I will. At least until the court decides.)
Last night, I got into a big political argument with a friend of mine. Probably one of my first real such arguments ever, as I tend to stay largely detached from politics. It wasn't actually an argument about politics, it was just one of those bordering-on-religious debates where both sides know they're right, and yet they have to try to convince the other of that, too. It's the most insidious kind of argument...even now I have this feeling that if I maybe just restate what I'm saying a certain way, then my friend would say, "Oh, that's what you mean, I totally agree," because it just seems so sensible. But I know that it doesn't work that way.
What touched it all off is partly a lawsuit that was filed recently, and partly a recent purchase of mine. The lawsuit was a class-action suit filed on behalf of computer hard-drive purchasers who were disappointed that when they stuck their hard drives in the computer, the computer told them that there wasn't as much space on them as they thought they were getting.
I read the article, thought, "Huh," but it wasn't quite brought home to me what it was really all about until I got a new hard drive for my Wallstreet laptop. The old one was flaking out, and I was constantly having to rebuild the directory just to get it to boot. Given that I could either spend $100 and get a 20 gig drive, or spend $30 or so more and get a bigger and faster one, I figured a 40 gig drive was worth the extra expenditure. I thought about how it would be nice to have those gigs...I could spend 8 on OS X, since it needs to be installed within the first 8 gigs of the drive on my type of computer, and then have 32 left over for mp3s or other miscellaneous stuff.
Only, when I got the drive and put it in the computer, it actually registered as having 37.26 gigabytes, not 40...which means my drive actually only has about 93% of the actual space on it that it claims to. The reason for this is that the hard drive manufacturers choose to use a different definition of "gigabyte" than the rest of us do.
You see, in computers, kilo means 1024. This is because computers think in binary, powers of 2. Thus, 210 = 1024...and that's close enough to 1000 that, needing a good shortcut, the computer hackers of the day named 210 bytes a kilobyte. Subsequently, 1024 kilobytes became a megabyte, and 1024 megabytes a gigabyte. This means that, by the definition commonly used in computer science, a gigabyte is 1,073,741,824 bytes. This is a very fundamental definition in the computer industry, and it's taught to even the most basic computer science student. But technically, it is also semantically incorrect.
This is because the prefixes kilo-, mega-, and giga- have specific meanings in the scientific community outside the computer field—they refer to 1,000, 1,000,000, and 1,000,000,000 respectively, and are commonly used to tack onto other units to specify 103, 106, or 109 of that particular unit: kilogram, kilometer, kilocalorie, kiloton, and so on. The difference between 1,000 and 1,024 really isn't all that great...but it increases geometrically as the powers increase...which means there's a difference between a gigabyte in the computer sense and a gigabyte in the standards sense of 73,741,824 bytes, or 70.33 megabytes, or about 6.87%.
So, why is there a 7% difference in the size that the drive claims to be on the package and the size it claims to be in the computer? Why, because the hard drive manufacturing industry chooses to use the 109 definition of "giga" instead of the 1,073,741,824 definition used by the entire rest of the industry! If you buy a 40 gigabyte drive, as I did, you're actually getting a drive that can hold 40,000,000,000 bytes, not a drive that can hold 40 true gigabytes.
70 megabytes per gigabyte may not seem like much—that's about 1/10 of a CD-ROM, 7/10 of a Zip-100 disk—but if you multiply it by 100, you begin to realize there's a pretty big difference there. And now that it's common for consumers to buy 150 or 200 gigabyte drives and get "short-changed" by 10 or 15 gigabytes, it's not too surprising they're suing. I'm a bit annoyed myself, and at myself, for not realizing it before I bought this 40 gig drive and lowering my expectations accordingly.
Anyway, the thrust of the argument my friend and I were having is that I feel the hard drive manufacturers are engaging in a deceptive practice, puffing up the size of their hard drives by using a different definition than everybody else in the industry in a manner not unlike candy bar manufacturers who put their bars in a package bigger than the bar itself. They're in the computer industry, there's no way they can avoid knowing how big the drive will claim to be in someone's computer. They should label their drives accordingly—consumers should have a reasonable expectation of getting a drive their computer agrees is 150 gigabytes when they buy a 150 gigabyte drive.
My friend, on the other hand, apparently feels that the drive manufacturers are using the correct definition of "kilo" and everybody else isn't—that using kilo to mean 1024 somehow dilutes the meaning of the standard and confuses people. He's not the only one to feel this way, of course; a standards group has come up with new names for kilobyte, megabyte, and gigabyte that they hope people will switch to using instead: kibibyte, mebibyte, and gibibyte. (Or, as I call them, kibble-byte, maybe-bite, and gibberish-byte.) My friend feels that in the court case, the hard drive manufacturers can bring expert witnesses forward and claim that they're using the correct definition of giga, and so it's okay.
I just don't think that's going to work too well, because the courts will look at the non-standard way in which kilo-, mega-, and giga- are used by the rest of the industry and decide that consumers should have a reasonable expectation that hardware manufacturers will follow the computer-technology (non-)standard when they're making computer parts.
Nor do I think it too likely that people will switch over to using kibi-, mebi-, and gibi-. Aside from the fact that they look and sound nonsensical, human nature is that when something erroneous or overcomplicated gets entrenched, it continues to be used through the ages even when people come up with something more correct or internally consistent. We're still using English measurements instead of the easier-to-convert metric system (though, granted, Europe's using metric more than we are). We're still on a 60 second minute, 60 minute hour, 24-hour (in two cycles of 12) day, 7 day week, 28/29/30/31-day month, 52-week or 12-month or 365.25-day year (and even that requires 'leap seconds' or skipped leap days every now and then) instead of coming up with something metric to replace it. We're still using the 19th-century QWERTY keyboard, which was designed to slow typists down so they wouldn't break their equipment, instead of Dvorak or one of the other systems designed to speed typists up. Kibi-, mebi-, and gibi- may be more accurate than kilo-, mega-, and giga-, but they came along about three decades of usage too late.
Unlike in French, which has an Academy to ride herd on the language, the meaning of words in English is created and changed by usage, not the other way around. So words and expressions that were originally incorrect become correct if enough people use them for long enough. It used to be a popular schoolyard expression to say, "Don't say 'ain't' 'cuz 'ain't' ain't a word!" But the expression itself is incorrect now, because "ain't" is enough of a word to have a lengthy definition in the Merriam-Webster Dictionary. Thus it is for kilobytes, megabytes, gigabytes, and so forth. But kibibytes and their ilk "ain't" in there yet.
Heck, it's not even the first time that "kilo" has been incorrectly used. Or, in this case, incorrectly not used. I learned back in my high school chemistry classes that there are actually two kinds of the unit of energy we call calories: calories and Calories with a capital C. The capital C Calories are actually 1,000 of the small-c calories, and are also called kilocalories. And guess which of these units is used to list the energy content of food? Kilocalories, or Calories. Except nobody but scientists calls them kilocalories, or bothers to capitalize when talking about "counting calories."
My friend feels that the courts aren't where this standards matter should be resolved...but I think they're the only place where the only part of it that can be resolved (getting the drive manufacturers to start labelling their products by the same definition everyone else uses) will be...and it had probably better be soon. The margin of error will only increase when they start selling terabyte (210 gigabytes) and petabyte (210 terabytes—the equivalent of twenty million four-drawer filing cabinets full of text) drives.
In the end, he and I simply had to agree to disagree, because there's no way we're going to be able to change each others' minds. (I still have this urge that pops up every so often to try it again, that surely he'll listen this time. But the wiser part of me says to just let it rest, and so I will. At least until the court decides.)
