Aarg! The kB(d) silliness is spreading to common shell commands. I’m not in hard drive marketing, so I want to keep using the standard kilobyte, not the ~2.34% smaller new decimal kilobyte. The spread of this silliness is almost like deciding that from today on electrons have a positive charge. Sure, it would make things clearer for people just getting started, but it causes an unnecessary ambiguity that just creates a new opportunity for mistakes and confusion in the future. Using prefixes that make sense in binary doesn’t follow the standard prefix pattern, but does use the closest pattern that makes sense in the number system being used. I think keeping the historical consistency and taking the base of the number system into account is more important than following a pattern established for a different base.

I’ll stick with the standard definitions:

  • 1 byte (B) == 8 bits (b)
  • 1 kilobyte (kB ) == 210 bytes = 1024 bytes
  • 1 megabyte (MB) == 220 bytes = 1048576 bytes

It’s getting harder to avoid the redefinition silliness. The only way I’ve found to fix it in OS X 10.6’s Finder is a hack. The change has made it into some common Linux distros, and even into the documentation for critical utilities such as rsync. Heck, I’ve even seen kB listed as KiB. This change even led to one recent textbook confusing the terms byte and word. Until this past year, any textbook or technical writing I had seen used the standard definition, with alternate terms used for any other similar terms. For example, 7- and 14-bit processors use different names. The decimal definition of kilobyte only appeared in marketing and other non-technical writing. So for consistency and clarity, I’ll stick with the traditional standard definition instead of the new decimal definition.