Since when are harddrives actually the size advertised?

October 2, 2009 4:07 AM

Since when are harddrives actually the size advertised? I just bought a Western Digital 500 GB 2.5'' HD and I was surprised when it actually turned out to be 499.32 GB after the first initialization (on a mac).

When has this changed? And what story/company was behind the adaption?

posted by mathiu to computers & internet (14 answers total)

Snow leopard counts bytes the same way the hard drive manufacturers do.
posted by motorcycles are jets at 4:13 AM on October 2, 2009 [3 favorites]

Firstly, there are two kinds of gigabyte, binary and decimal. The former is one billion bytes (which makes sense really in a way, since we mostly use decimals). The latter results from the fact that we use 1024 bytes for a kilobyte (which kind of makes kilobyte a misnomer). A gigabyte, similarly, is 1,024 Megabytes or 1,073,741,824 Bytes.

That would have got you 465GB in Windows, because Windows uses base 2 to calculate disk space. Your Mac (as m.a.j. explained) is using base 10, which gets you much closer to the 500Gb you're expecting.

Secondly, there's unformatted and formatted capacity. Your disk, in its completely empty state, has 500GB of space. Your operating system, however, needs somewhere to keep its file allocation table (an index which tells it where to look for files), and this eats up some of the space when you format the disk. The remainder is available for storing files, so that's what you see.
posted by le morte de bea arthur at 4:17 AM on October 2, 2009 [6 favorites]

Hard disks have always been the sizes advertised. An advertiser's gigabyte is almost always 1,000,000,000 bytes. Historically, an engineer's gigabyte has usually been 1,073,741,824 bytes when talking about data storage, or 1,000,000,000 bytes when talking about data transmission. Yes, this is deeply unsatisfactory, and has been recognized as such for many years.

Interestingly, flash disk capacities are still generally advertised in engineer's gigabytes (which really ought to be called "gibibytes", and probably would be if "gibibyte" were not such a dorky sounding word). So if you stick an 8GB USB stick in your Snow Leopard Mac, you'll probably find that it's magically grown to 8.6GB :-)
posted by flabdablet at 4:39 AM on October 2, 2009 [1 favorite]

Since when are harddrives actually the size advertised?

Never have been, never will be. The discrepancy between advertising (decimal) sizes and proper (binary) sizes has existed since before 5 1/4" floppies - it's just that now that drives have reached terabytes sizes, the discrepancy is much more noticable.

Technically, you should call the binary gig (2^30 bytes), which your OS will report on, a gibibyte, and a decimal gig a gigabyte (10^9 bytes)... but no-one really does that.

People selling hard disks use the measure that makes their disk look better, and always will.
posted by pompomtom at 4:55 AM on October 2, 2009

It's worth noting that the discrepancy between 'advertiser' and 'engineer' units is getting worse with time. Back in the day, we cared about kilobytes. The difference between a binary kilobyte (1024) and a decimal kilobyte (1000) is only 24 bytes, a measly 2.5%. Now, we care about terabytes. A binary terabyte is 1.0995 x 10^12 bytes, and a decimal terabyte is only 1 x 10^12 bytes, a whopping 10% difference.

Since the problem will carry on getting worse, it's high time for standardisation. As a scientist, I'm glad that Apple chose to support the International System of Units (SI), which is a very powerful tool.
posted by beniamino at 5:15 AM on October 2, 2009

erm 2.4%, obviously
posted by beniamino at 5:16 AM on October 2, 2009

It is also important to note that historically computing used the binary gig (a more natural measure in that context). Thus 1 GB really was 2^30.

It wasn't until the hard drive manufacturers started to mis-represent the size of drives using the decimal gig that the new standard described by pompomtom needed to be put in place. This is relatively recent (20 years maybe?) in the grand scheme of things.

I believe there were several law suits when this confusion first came to the mainstream attention.
posted by NoDef at 5:18 AM on October 2, 2009

Even though Snow Leopard reports the drive as 500 GB when Leopard (or Windows) reports it as 465 "GB" (really GiB), it's still the exact same size with the exact same amount of usable space.

It would be like if there were no word for "liters" and people in Europe used the word "gallons" differently than people in the U.S. use "gallons." If we were both looking at the same jug of milk, the Americans would say it's 1 gallon and the Europeans would say it's 3.78 gallons. ("That's preposterous," you might say, "why would people use the same word to refer to two completely different ways of measuring the same thing?" Well -- that's exactly how bytes have been measured these past few decades.)
posted by Nothlit at 7:09 AM on October 2, 2009

It would be like if there were no word for "liters" and people in Europe used the word "gallons" differently than people in the U.S. use "gallons."

A British gallon (rarely used now) is 20% larger than an American gallon.
A British pint (only used for milk bottles and beer these days) is 20% larger than an American pint.
A metric ton is slightly lighter than a British ton which is 10% lighter than an American ton.
posted by atrazine at 7:25 AM on October 2, 2009 [2 favorites]

To be fair, the British ton and gallon have been out of popular use for quite a while now (thankfully).
posted by le morte de bea arthur at 7:48 AM on October 2, 2009

It depends how you define "popular use" - my parent's generation still use the term gallon and ton in their original forms plenty (at least, my parents and my in-laws do - but my in-law, a carpenter, can only understand millimetres and not centimetres or metres..).
posted by wackybrit at 8:41 AM on October 2, 2009

Its also important to consider the filesystem the disk is using. Only x amount of space is actually usable as the rest is used by the filesystem scheme.
posted by damn dirty ape at 10:10 AM on October 2, 2009

flabdablet wrote "Interestingly, flash disk capacities are still generally advertised in engineer's gigabytes"

Wow. I didn't notice that - but yep, my 16gb USB is reporting as 16.17 in Snow Leopard now. Interesting.
posted by caution live frogs at 12:12 PM on October 2, 2009

It wasn't until the hard drive manufacturers started to mis-represent the size of drives using the decimal gig...

PC hard drive manufacturers have been using the SI, as opposed to compsci, prefix meanings since well before the first gigabyte-capacity drive was a glimpse on the horizon. The seminal Seagate ST506 drive had a formatted capacity of 153 cylinders of four tracks of 32 sectors of 256 bytes = 5013504 bytes, and was marketed as a 5MB drive. A 5MiB drive with the same track formatting would have required 160 cylinders.

I worked for a company that made products based on those drives. By using a controller capable of handling 512-byte sectors instead of 256-byte sectors, we were able to fit 17 sectors per track. 153x4x17x512 bytes = 5326848 bytes, which actually is a smidgen over 5MiB.

We were really proud of our hard drive unit's transfer rate, too. By using careful sector interleaving and 1 megabyte/second DMA, we could achieve over 90KB/second file read speed in Apple II ProDOS.

Just the other day, I happened to measure the sustained read speed on the drive in the box I'm using right now. It's well over 1000 times faster. At 1.5TB, it's also nearly 300,000 times bigger. Kids these days...
posted by flabdablet at 9:38 PM on October 2, 2009