I don’t argue you’re definitions, but it seems to me that not all kilos, gigs, megas, etc are exact. They vary on “approximation”, based on what you are saying, which would make metrics not as great as some people make it out to be. Thus, I can go in one place and a kilo will be a 1000, but I go somewhere else and it could be 1024 and yet another place 1050. The same with the others. Those numbers seem to vary all the time, based on human approximation. So you contend that it is 1000, but I go in to look at a computer and they contend it is 1024. I use a computer a lot and those are the numbers I have to use, if I remember those numbers, when I talk to a geek or a nerd, if we use numbers at all for what we are discussing. However, as GdB showed, it’s pretty standard.
The short answer is: yes, the size of harddrives and disks were simply an approximation, the sizes that were actually in powers of two, where approximated as sizes of powers of 10, and so the box labels had merely approximate numbers, always. Sometimes software used the base 10 approximation, sometimes they used the actual base 2 sizes. When the sizes were small, the approximations were easier to discuss.