One bourbon, one scotch, and one mav

One of the things I hear at least every day on the phone is the formatted space fallacy. SCSI disks come in fairly fixed capacities. So when I’m replacing a drive, I ask something like “Is that drive a 73GB hard drive?” The customer invariably answers either “yes” if they do this shit for a living, or “It’s like 68,349KB” or some shit like that if they don’t. Then, usually, the latter person will say something like “Oh, that’s the formatted capacity.”

They’re only half right. Disk operations (partitioning, formatting, etc) do consume space on the disk. However, the space used by the filesystem is generally substantially smaller than the space “lost” by numeric conversion. Why? For a really long time storage has been rated in the metric version of the binary prefix in question.

“What in the ever-living fuck does that mean?” I can hear you asking already. So we’ll take a look at an example.

The prefix “kilo” represents 10. So in a computer, a kilobyte would normally reference a size of 2^10 or 1024. So a 4KB cluster would store 4096 bytes. However, in the world of storage (and the more generic world at large) represents 10^3, or 1000. So in a hard drive, 4KB would be 4000, a “loss” of 96 bytes. In reality, nothing has been lost. However, as most computers report EVERYTHING using the base-2 idea of measurement, not the base 10 idea, your hard drive appears to be smaller.

Real world example: The hard drive in the Mac I am posting this from has 100,030,242,816 bytes.
A hard drive manufacturer would reference this roughly as a 100GB hard drive.
The detailed specs would say the hard drive has 100 gigabytes, and they’re not lying: if we shift the decimal point 9 spaces right (as a gigabyte is defined in base-10 as 10^9) we get 100.030 GB. However, the computer I’m at says it is 93.2 GB. If you’re calculating a gigabyte as 2^30 instead, this is also perfectly accurate. It’s far short of 100GB by base-2 standards, as 2^3o = 1,073,741,824 bytes.

There is a certain faction within the industry that thinks this tactic is dishonest. Is this ripping people off? Well, not really. Measurement is pretty abstract, and there are many industries that use multiple types of measurement.. When it comes right down to it, every measurement is just an agreed-upon standard. So, if the storage biz wants to agree on using base-10 measurement for their storage devices, who cares?

But the least we could do is to not be fucking ignorant about it.

For reference:
Binary prefixes
Manufacturer’s size vs. reported capacity