Apple Kicks HDD Marketing Debate Into High Gear

quacking duck writes "With the release of Mac OS X 10.6 Snow Leopard, Apple has updated a support document describing how their new operating system reports capacities of hard drives and other media. It has sided with hard drive makers, who for years have advertised capacities as '1 GB = 1,000,000,000 bytes' instead of the traditional computer science definition, and in so doing has kicked the debate between marketing and computer science into high gear. Binary prefixes for binary units (e.g. GiB for 'gibibyte') have been promoted by the International Electrotechnical Commission and endorsed by IEEE and other standards organizations, but to date there's been limited acceptance (though manufacturers have wholeheartedly accepted the 'new' definitions for GB and TB). Is Apple's move the first major step in forcing computer science to adopt the more awkward binary prefixes, breaking decades of accepted (if technically inaccurate) usage of SI prefixes?"

Comments

Its been done for years already

[sopssa]

Is Apple's move the first major step in forcing computer science to adopt the more awkward binary prefixes, breaking decades of accepted (if technically inaccurate) usage of SI prefixes?

No, its not any first major step. HDD makers already went there years ago, its established and people know better what it means. And even if I'm quite a nerd myself, I never think that 1 terabytes = 1 048 576 megabytes. Yeah it would be great if I remembered that or as many decimals in PI as possible, but no one really cares. It's a lot easier to remember and think that 1 terabyte is 1 000 000 megabytes, even if its not technically so because of binary system and even if I know that - I still think so just for the easy of it.

And its a mac. What did you think? It's as far from a nerdy computer as possible. Obviously they are going to use terms and units that non-geeky people understand.

Re:Its been done for years already

[schmidt349]

What's more, Apple's been sued a couple of times over the definition of a gigabyte by angry idiots who didn't understand that 10^9 != 2^30. Possibly they're doing this in part to minimize their future liability.

Re:Its been done for years already

[Anonymous Coward]

I'm quite a nerd myself, I never think that 1 terabytes = 1 048 576 megabytes

Well of course not; why the fuck would you want to? That's like wondering how many hours there are in a week - who cares? 1 terabyte is 1024 gigbytes. Converting it into megabytes is pointless for the purposes of most people.

Hey let's have a 10 bit byte as well to make conversions that nobody ever does, easier.

Re:Its been done for years already

[sopssa]

Your example is bad because its the default one. 1 terabyte to 1024 gigabytes is easy. How quickly you calculate that to 4TB? 15TB? 492TB? Or for more better example, 405GB to MB's? Its just a lot easier to think 405GB = 405 000MB than start calculating it, while its kinda close anyway.

Re:Its been done for years already

[hedwards]

Not really, how many hours in a week is a lot easier to do in your head than how many bites in a terabyte. Additionally, the computer scientists shouldn't have been using prefixes that already had a meaning.

And BTW, the answer is 168.

Re:Its been done for years already

[Kokuyo]

Seriously, did you ever need to? I've been in IT since 1998 and I cannot remember ONE situation where I thought "This is so inconvenient, I need a calculator for this shit. Couldn't they just make a Gigabyte 1000'000'000 Bytes?"

So we've had a defined standard that was, arguably, not the easiest to understand. THEN harddrive manufacturers started their fraud. And THEN people started complaining. So what, and please think about this, would be the right decision here?

As to being complicated: If that is your argument, then all the English speaking countries should switch to metric according to your logic. Obviously, a lot of people don't like that. So why is it okay here and not okay there?

Re:Its been done for years already

[mwvdlee]

The problem isn't the definition, it's that OS's and hardware manufacturers have been using different definitions. If both of them would stick to factors of 1000, there would be no problem. If they all stick to 1024, there would be no problem. The problem is that both definitions are used.

Personally I'd vote for 1000, since it's just easier for most people. That way they could easily know that 1001 1MB files do not fit on a 1GB USB stick and all the world would be consistent.

Re:Its been done for years already

[mftb]

Because as far as disk space occupation goes, that file may as well be 16KB.

Re:Its been done for years already

[mokus000]

A bigger issue, for me, is why the stupid Finder reports file sizes based on blocks! This makes no sense. I can plug in a flash drive, and the Finder will report that a 12KB file, copied to the desktop, is now a 16KB file. This isn't rocket science, FIX IT already, Apple!!

Well, given an 8k or 16k block size, a 12k file *DOES* consume 16k of usable disk space. Plus 600-700 bytes for the inode and directory entry. Plus more if there's any magic Apple-y metadata associated with the file.

For what reason do you expect any filesystem browser to report the exact number of bytes in a file? I'm almost always more interested in knowing how much disk space is used by the file - 16k in your example. In a filesystem like JFS that dynamically allocates inodes, I might even expect it to report the space used by the inode. FWIW, 'du' will report 16k in your example as well. Is 'du' wrong too?

Also, what should it report for directories? Taking a directory of the source of GHC 6.10.4 on my computer as an example (lots and lots of smallish source code files):

$ find . -type f -exec cat {} \; | wc -c
  29776950
$ du -sk .
35036 .

Those numbers don't match (taking into account the conversion between bytes in the first case and kb in the second), but I can't see a reason ever to care about the first one. It's not even a very good indicator of what size an uncompressed tar file would be.

Finally, I just went and took a look at a small file on the desktop of my mac. "Get Info" tells me:

Size: 8 KB on disk (782 bytes)

So it *does* report the number of bytes in the file, as well as the disk usage, clearly labeled. Now I really don't exactly know what you're whining about.

Re:Its been done for years already

[Anonymous Coward]

then all the English speaking countries should switch to metric according to your logic.

Yes. Yes they should all switch to metric.

Re:Its been done for years already

[SirCowMan]

'du', disk use, obviously should describe the actual used space on the drive, as that is the name of the program. I, however, would rather any other form of file management to note the physical size of the data in the file. Checking file sizes against, say, a website you just uploaded is a quick and easy way to ensure it all transferred for example.

Re:Its been done for years already

[mfnickster]

>> Now I really don't exactly know what you're whining about.

> Because it's Apple. Are you new here?

Actually, you're kind of right there. Apple was at the forefront of making computers usable for the average person, and their Human Interface Guidelines specifically recommend that the computer be made to work the way people do, rather than making people work the way the computer does! :)

Re:Its been done for years already

[rolfwind]

The problem isn't the definition, it's that OS's and hardware manufacturers have been using different definitions. If both of them would stick to factors of 1000, there would be no problem. If they all stick to 1024, there would be no problem. The problem is that both definitions are used.

The problem is precisely the definition, or rather that computer people think messing with "mega", "kilo", etc is okay because it's their own niche. Mega is understood as 1,000,000 and kilo as 1,000. I got a CS degree, and I always thought it was stupid how we subverted the meaning. 2^10 aka 1,024 is arbitrary, is in no way 1000 and was chosen purely because it was the closest power of 2 close to 1,000. What if every niche started subverting commonly understood scientific measurements for their own convenience?

We defined bit and byte and the like. Great. We could do that. But we should have left mega and all the prefixes alone. If we weren't happy wit that, go with our own, like 'mebi' series of prefixes has attempted.

Re:Its been done for years already

[norton_I]

So we've had a defined standard that was, arguably, not the easiest to understand. THEN harddrive manufacturers started their fraud. And THEN people started complaining. So what, and please think about this, would be the right decision here?

This is revisionist at best and really just wrong. Despite all "wisdom" to the contrary, there has never been a universal acceptance of 1 MB = 2^20 bytes on computers. For instance, all of IBMs mainframe hard drives from the 60s and 70s were sold using base-10 prefixes. Early desktop hard drives from the 80s used both. I think the ST506 used base-2, but some other models used base-10. All networking and communications standards (ethernet, modems, PCI, SATA...) use base 10 prefixes for MB/s and Mbit/s. 3.5" floppy disks used NASA-style units where 1 MB = 10^3*2^10. Even while RAM is still almost always measured in base-2 units (due to manufacturing issues making it much easier to produce in power-of-2 sizes -- something which is not true for hard drives) the speed of the memory bus on your CPU is still measured in base-10 units.

It is a *good* idea to have K and M mean the same thing everywhere. A system where a 1 GB/s link transfers 0.93 GB every second is stupid. This is especially important as computers are being used in more and more environments. Should a 1 megapixel camera mean 2^20 pixels? What about CDs with a 44.1 KHz sampling rate?

Re:Its been done for years already

[mfnickster]

> 1024 is NOT arbitrary.
> ...
> Because computers work in powers of 2.

What he means is, it's an arbitrary choice of *grouping* - there's nothing in the base 2 or base 10 systems that puts 1024 on a digit boundary.

1024 is 2^10 - to be self-consistent, they should have chosen 2^8 or 2^16 for grouping, since 8 = 2^3 and 16 = 2^4, but they chose 2^10 because it happened to be "close to 1000"

They took the "kilo" prefix out of convenience and wedged it into a system not suited for it.

Re:Its been done for years already

[Kjella]

Because computers work in powers of 2.

Except when they don't, like floppies, CDs, DVDs, BluRays, HDDs, dial-ups speeds, networks or any of the many other places where they don't. Eventually you run into issues where there's a GigE (1,000,000,000) network adapter running a 3GHz (3,000,000) processor which is processed in 512 MB (512*1024*1024) RAM before being stored over a 3 Gbit (3,000,000,000) SATAII connection to a 1TB (1,000,000,000,000) hard disk. Every time you run into other sciences like "we need to process 1000 samples/second at 16 bits, that's 16 kbits right?" you run into trouble.

On the other hand, I can go into the details and say that in order to fit the CPU L1 cache it's 64 kB (64*1024) and textures can be maximum 2048*2048 pixels and there are exactly 512 stream processors to work with, you can handle 2^32 bits in an integer and so on and so forth. We're never going to get to where we can ditch base 2 sizes either, they're vital on almost every level once you get into the details.

Everytime you say "this is not a problem, because computers don't interact with the rest of the world and/or it's always trivial to tell" you are seriously deluding yourself. All the people saying "you should all use kB = 1000 and forget the rest" or "you should all use kB = 1024 and forget the rest" are both deluding themselves. We need both and we need clearly defined units for both. That's why I now say use kB = 1000 where it's correct. "Losing" the battle over kB is the only way we'll have kB and KiB, because clearly it's impossible to change the meaning of kilo = 1000 in everything else.

Is that why

[MeNeXT]

snow leopard frees 7gigs? Because it can't do the math? #8^)

Re:Is that why

[TheRaven64]

The Darwin versions of utilities like du and df have had the -h and -H (human readable numbers with either binary or decimal prefixes) the opposite way around to FreeBSD since 10.5. They made the existing switches, that had always reported the power-of-two sizes, display the power-of-ten ones and moved the old behaviour to the new option. In FreeBSD, they added new options for the power-of-ten versions. I wondered why my files suddenly became smaller after copying them to a FreeBSD machine for a while before I noticed this.

Re:Is that why

[ljaguar]

how did this get modded up? this is misinformation.

du(1) man page (snow leopard):
          -H Symbolic links on the command line are followed, symbolic links
                          in file hierarchies are not followed.

          -h "Human-readable" output. Use unit suffixes: Byte, Kilobyte,
                          Megabyte, Gigabyte, Terabyte and Petabyte.

df(1) man page (snow leopard):
          -H "Human-readable" output. Use unit suffixes: Byte, Kilobyte,
                          Megabyte, Gigabyte, Terabyte and Petabyte in order to reduce the
                          number of digits to three or less using base 10 for sizes.

          -h "Human-readable" output. Use unit suffixes: Byte, Kilobyte,
                          Megabyte, Gigabyte, Terabyte and Petabyte in order to reduce the
                          number of digits to three or less using base 2 for sizes.

this is exactly same output as man pages fro those two in FreeBSD 6.1

this is man page from debian linux:
              -h, --human-readable
                            print sizes in human readable format (e.g., 1K 234M 2G)

              -H, --si
                            likewise, but use powers of 1000 not 1024

so it seems to me that behavior of darwin is exactly same as gnu tools.

Tilting at windmills

[Melkhior]

The SI prefixes have been around for nearly 5 decades, and have a specific meaning used by everybody. Every scientist uses them in one way or another, and for every last one of of them, they have the same meaning.

Why can't we, the C.S. people, accept that?

Giga is 10^9. It has been 10^9 since it was created. It was never, ever meant to be anything but 10^9.

If you want to talk about 1024^3, then it's Gibi. Gibi is 2^30 since it was created. It was never, ever meant to be anything but 2^30.

Get over it.

(and yes, I try to always use GiB whenever it's appropriate).

I agree with HD manufacturers too.

[Speare]

For things where there's a clear "address bus" that consists of all possible permutations of a binary bit field, it makes sense to use the powers of two. The 2^10 kilo-, 2^20- mega, 2^30 giga- is just a convenience in terminology due to their approximate equivalence to 10^3, 10^6, 10^9, respectively; however, the bigger you go, obviously they diverge quite a bit.

For things addressed by a system of arbitrary track/cylinder numbers, say, 336 tracks or 1435 tracks, and arbitrary platter/head numbers, it's ridiculous to say that they should follow the "convenience" of the powers of two scheme.

So, how should flash drives be measured and marketed? While the components are physically based on an address bus, they present themselves to the host with sector numbers just like the spinning drives do. They can also reserve some "spare" cells in their internal mapping, for wear-leveling or error correction. I'd say they could easily make the case for marketing under SI/IEEE powers of ten.

1 MB = 1,000,000 bytes

[moon3]

http://en.wikipedia.org/wiki/IEEE_1541

These IEEE recomendations seam like common sense to me.

1 KB = 1,000 bytes
1 MB = 1,000,000 bytes
1 GB = 1,000,000,000 bytes


And for you droids and androids out there:

1 KiB = 1,024 bytes
1 MiB = 1,048,576 bytes
1 GiB = 1,073,741,824 bytes

Re:1 MB = 1,000,000 bytes

[smaddox]

If it is more convenient to use KiB, MiB, and GiB, then use them. Just don't call them KB, MB, and GB.

I don't understand why the hell this is so difficult.

The ripoff scales non-linearly with size

[Ant P.]

For 1KB the difference is 24 bytes.
For 1MB, 2**20 - 10**6 = 48576, 48KB difference or 4.6% less than the larger of the two.
For 1GB, 2**30 - 10**9 = 73741824 (73MB), 6.9%.

For a 1TB hard disk you're being short-changed by 9%: 94 gigabytes!

Re:Okay, so technically,

[TheCount22]

Kilo = 1<<10
Mega = 1<<20
Giga = 1<<30
Tera = 1<<40

Goes by ten!

Use the standard terms in the standard way

[cheebie]

As much as techies complain about people using technical terms inaccurately, we should use the SI prefixes in ways that mean what they mean. The fact that 2^10 is close to 1000 doesn't mean we get to hijack K/M/G to mean 2^10/2^20/2^30.

And mentally we're using them to mean powers of 1000 anyway. How often do you _really_ mean 1024 when you say 1K? Personally, I'm always thinking 1000-ish.

SI units vs. binary units

[Omnifarious]

I'm happy Apple is doing this. The use of SI unit names for base 2 values was convenient and gave relatively small errors for low numbers. But up above a gigabyte, and certainly in the terabyte range it's just plain wrong. And certainly nobody who's not a CS person is going to think "Oh, yeah, I divide the base 10 exponent by 3 and multiply by 10 to get the base 2 exponent because this is a piece of computer equipment!".

The binary SI prefixes aren't that hard to use when they really make sense. Computer science should get with the rest of the world in how things are measured and quanitifed and stop doing so with its own special language understood by those well versed in the field unless that language uses words and terms clearly different from the standard ones.

Re:bug

[RedK]

Yeah, but the alternative Mebi and Gibi sounds like something out yaoi. So I'd rather stick with 1 Gigabyte = 1024 Megabytes.

Tebi, Zebi, and Pebi?

[Sububer]

Seriously? These sound like next generation Valley Girl names, not self-respecting geek prefixes.

When using prefixes that end in 'a' or 'o', I feel macho. Megabyes! Teraflops! Yottapwnage! Yeah, baby!

From my cold, dead hands, Apple.

BTW, who thought of the cutsey name "Apple" anyway? Nice name. Pfft.

that "standard"?

[Mathinker]

The "standard"? All of the standards associations recommend using G/M/K as prefixes with the base-10 meanings, and using the unambiguous Gi/Mi/Ki (gibi/mebi/kibi) for base-2 measurements. One standards organization was willing to allow the deprecated use of G/M/K as base-2 for measuring semiconductor memory (i.e., RAM) only.

Do you also recommend that we will suddenly measure disk drive capacity in a different unit if/when we all move to using quantum computers or computers based on some other new currently unfamiliar technology?

Oh, and BTW, at least one of the technologies which has a small chance of replacing current RAM technologies, phase-change memory, could theoretically store 3 or 5 states per unit cell instead of 2 or 4, given the right material undergoing the phase change. One of the reasons not to do it is because it would be a pain to convert to and from base-2 to interface with the computer, so in the long run it is possible (but not necessarily likely, because there is a large initial development cost) that some computing devices will be designed to work in base-3 or base-5 if only to better utilize the abilities of PCM.

Re:I have a better idea...

[HouseOfMisterE]

The negative terminal of a battery supplies the electrons and they move from negative to positive when a conductor is placed between the two poles. The two popular notations for charge flow, "Conventional Flow Notation" and "Electron Flow Notation", do not dispute this. The difference is that "Electron Flow Notation" illustrates the physical movement of electrons (from "negative" to "positive") and "Conventional Flow Notation" illustrates the "movement" of the electrical charge from the "positive" terminal to the "negative" terminal. As electrons move from - to +, the "positive" side of the battery becomes less positive in relation to the "negative" side, effectively meaning that the electrical charge is moving from + to - (in "Conventional Flow Notation"). The electrons are still moving from the - battery terminal to the + battery terminal, though.

Re:There's a debate? Don't think so

[zzatz]

"This is a case of marketing trumping computer science."

No, this is a case of standards trumping common (mis)usage. Metric prefixes have been in use for centuries, and they are powers of ten. That's how the national and international standards have ALWAYS used them.

Those prefixes are convenient, and have been used for powers of two in casual, informal usage. But powers of two were never part of any official standard until recently, when NEW and DIFFERENT prefixes were added.

Scientists and engineers have always used powers of ten. Manufacturers used to be careful to distinguish between the formal definition (powers of ten) and the casual usage (powers of two). For example, Intel lists the exact number of bytes in parentheses whenever they use the casual meaning of the prefixes, showing that they were aware of the potential for confusion.

But many reporters and hobbyists were not trained in engineering or science, and missed the distinction. So you ended up with what I think of as "AOL prefixes". Microsoft ignored the standards, as they so often do. They may have been confused by earlier systems, such as UNIX and RT-11, which reported space in numbers of disk blocks, rather than bytes. In early UNIX, the ls command lists the number of bytes without prefixes, and the du and df commands list the number of disk blocks, not the number of bytes.

I don't expect hobbyists or journalists to get the prefixes right. I can live with the misuse of the prefixes. But it really bothers me when someone complains when the prefixes are used correctly, in compliance to published international standards.

OSX does this too...

[PC and Sony Fanboy]

If you right click a file in Windows and go to Properties you see:


Size: 2.47 KB (2,539 bytes)
Size on disk: 4.00 KB (4,096 bytes)


I thought Mac OS X was supposed to be easy?

Mac OS X does this as well.

The problem is that mac users don't know how to use a computer ... ergo, they are mac users.

Re:Benchmarks

[zippthorne]

No they won't. Network speeds use the same terminology that RF engineers use: base-10 prefixes.

Re:customer enlightenment and its drawbacks

[Melkhior]

Uh.. the inch is technically an SI unit. It is defined as exactly 2.54 cm.

No, it's not. SI uses the metre for length measurement, and nothing else. You can alter it with the various prefixes, and there's is only one thousand meters in a kilometre, not twenty-four more.

The "inch" from the United States customary units is defined as 2.54 centimetres, but it doesn't make it part of the SI..

Re:I honestly don't care much whether I'm getting

[Skreems]

Now granted, the actual disk capacity hasn't changed by a single bit as a result of changing the notation from a mislabeled TiB to actual base-10 TB, but it at least makes buying the biggest, most expensive drives a little less painful since they don't appear 10% smaller right out of the box.

I've got bad news for you... while your drives no longer appear 10% smaller, all your files are now 10% larger.