Ask a physicist and a 1990s programmer what a kilobyte is, and you will get two confident, incompatible answers: 1000 bytes and 1024 bytes. Both have history on their side. The result is one of the longest-running ambiguities in computing, one that still shows up every time a new hard drive looks smaller than the box promised.
The root of the problem is a numerical coincidence. Computers organize memory in powers of two, and 2 to the 10th power is 1024 — temptingly close to 1000, the value the metric prefix “kilo” has meant for two centuries.
How the collision happened
Early programmers, needing a word for 1024 bytes, borrowed “kilobyte” because it was only 2.4 percent off. The borrowed meaning worked fine while quantities were small, but the error compounds with each prefix step: a binary megabyte (1024 squared) is about 4.9 percent bigger than a metric one, a binary gigabyte about 7.4 percent, and a binary terabyte nearly 10 percent. What began as a rounding-friendly pun grew into gigabytes of genuine discrepancy — enough to spark consumer lawsuits against drive makers, who used the decimal definition while operating systems reported in binary.
The official fix nobody uses
In 1998 the International Electrotechnical Commission introduced dedicated binary prefixes: kibi (Ki), mebi (Mi), gibi (Gi), and tebi (Ti), so 1 KiB is exactly 1024 bytes and 1 kB is exactly 1000. It is a clean solution, and technical standards, Linux tools, and some software adopted it. But mainstream culture largely did not. Windows still reports file sizes in binary units while labeling them KB, MB, and GB; macOS switched to decimal units years ago with the same labels. So the identical file can show different “megabyte” sizes on two computers sitting side by side, both claiming to use the same unit.
Surviving the ambiguity
In practice, context usually tells you which definition is in play. Storage marketing, network speeds, and DVD-era optical media use decimal prefixes. RAM is always binary — a 16 GB memory module really contains 16 GiB. File sizes depend on the operating system displaying them. When precision actually matters — capacity planning, billing, scientific work — the safe move is to state sizes in exact bytes or use the unambiguous KiB and MiB notation, letting no reader guess.
The practical takeaway: treat “kilobyte” as a word that needs context, assume decimal for disks and networks, binary for RAM, and reach for the explicit KiB/MiB prefixes whenever a misunderstanding would cost you something.
Loading comments…