Fransisc H is the only one who's had it right.
Increments of 1000 are used by the industry to make their hard drives and disks seem bigger than they are (like selling DVD's that are "4.7GB" when, in fact, they have only 4.35 GB of space available on your PC)
Bytes are based on bits, using a "binary" system (base-2); one byte is 8 bits (2^3), one kilobyte is 1024 bytes (2^10), one megabyte = 1024 kilobytes (2^20), etc.
Resellers use the decimal (base-10) system:
One kilobyte = 1000 bytes (10^3), one megabyte = 1000 kilobytes (10^6), etc.
Using this system, numbers grow bigger faster in comparison... but in fact signify LESS data than under the binary system.
On a computer, when dealing with file- and drive sizes, the base-2 or binary system is used - that's simply how they're setup.