In the last few years binary prefixes have started to gain traction, with some operating systems picking them up and applications or storage requirements making use of them. The problem is that they are ridiculous sounding and completely unnecessary. The “problem” is that the standard better known SI units (mega, kilo, giga etc) are based on powers of 1000, while computer units are based on powers of 2. A good example of this is hard drives, where a 20GB HDD would be expected to be 21474836480 bytes but in reality would be 20000000000 bytes, losing 1474836480 bytes or 1.3GB,. For a long time people were angry at HDD manufacturers for deceiving people or falsely advertising the HDD capacities, although recently with binary prefixes being adopted people have started to defend manufacturers.
Well, the people who are defending them are simply wrong. It doesn’t matter that binary prefixes are technically correct. For over a decade before they were adopted, SI unit prefixes used in computing has constant standardized meanings that were accepted everywhere by everyone. It was HDD manufacturers who decided to ignore this and redefine SI units used in computer to match their more standard definition to save money. Not to make it easier for consumers, simply to save money. It didn’t make sense then, and it doesn’t make sense now. No other product related to computing uses the general meaning of SI prefixes when applied to computer units. HDD capacities should always be based around powers of 2, not powers of 1000 simply because that’s what computers use.
The reason SI units are not a problem in computing most of the time? Context. It is clear that the definitions are slightly different in computing and it isn’t hard to learn the differences. It certainly makes sense to have a separate and subsequently more accurate standard prefix specific to computing, but that should have been implemented earlier on. As things currently stands the name are dumb (kibi, mebi, gibi), they introduce confusion since people have already adapted to the existing system and there isn’t really a need for them because of that. Despite many people being opposed to them and better solutions existing (such as Knuth’s large kilobyte) it seems binary prefixes are here to stay, with the EU mandating their use in law. It will be interesting to see if they are prevalent in 10 years or so or if they rightfully vanish into the nether.