in case of bytes, powers of two actually make sense.
Absolutely. And the scheme that I outline takes advantage of that happy mnemonic near-coincidence. Without any drawbacks. And eliminates the ambiguity that has led to fuckup after fuckup after fuckup after...
Whoever later "decided" that 1 KB = 1000 bytes was a moron
Nobody later decided that - the prefix 'k' or 'kilo' has denoted exactly 1000
since long before the dawn of logic-based computing.
It's not used by anyone.
It is used by every thinking person that values precision.
Universities don't teach it that way.
Universities absolutely teach the difference between 1000 and 1024, and the difference between metric powers-of-ten units and binary prefix powers-of-two units. Sure, there may be some backwaters that are still stuck in the eighties*, clinging to a misleading and imprecise anachronism. But by and large, the difference is known and taught - everywhere that values truth.
*Actually, I graduated with an Electrical Engineering degree in the eighties. And while we didn't yet have the convenient binary prefix notation back then, we were
absolutely taught that misusing metric prefixes as an approximation was exactly that - an
approximation. And
not strictly correct. And to always be mindful of the error and ambiguity inserted due to this misuse.
Your operating system (whenever if it's windows or linux) doesn't show it that way.
Linux
certainly knows how to report memory and disk capacities in KiB, MiB, GiB, TiB, ... Of course, if your favorite clownshoe hides the truth from you, don't put that on Linux - put it on your dumbed down distro. And don't talk to me about Windows as if it is some paragon of Vires in Numeris. Last time I bothered to look, Windows couldn't even agree with itself. What with it reporting a GB as (something like) 1024*1024*1000 in Disk Mangler and as 1024*1000*1000 in File Mangler. Or some equivalent comedy of hapless errors.
1 KB = 1000 bytes is only in the head of some senile bureaucrats that don't understand technology and memory alignment.
Hoo-boy. we got a live one here. I suppose you think NIST is ignorant on computing? ISO? IEEE? ACM? IETF?
Just because a bureaucrat says the earth is flat it doesn't mean it is.
Cool. You've said a second thing that isn't 100% wrong. But conversely, just because
you say that 1000=1024 does not mean it
is.
You can't wish it that way, just like you can't make a memory chip that holds exactly 1000 bytes of information.
This has exactly zero to do with the discussion at hand. However, this could be good for a couple of laughs. Please explain to me the physics behind your assertion that one
cannot build a memory chip that holds exactly 1000 bytes of information. On second thought, don't. We don't need the fruitless diversion. Nor the ensuing hilarity.
So now that that is out of the way, let me try again. If we use the binary prefixes when we are speaking of 2^(10*n):
- we remove the ambiguity of not knowing if a measurement of (e.g.) 'M' means 1000000 or if it means 1,048,576 (or even 1,024,000);
- we lose none of the brevity of compact notation;
- we lose none of the mental chunking that comes from conveniently scaled units;
- we stop losing time and money due to fuckups in unit misunderstandings;
- we are using the units that all serious and conscientious scientists and engineers use;
- we stop killing people; and
- when the legacy Bitcoin max block size is 1MB, we know instantly it is 1,000,000 bytes, not 1000*1024 bytes, nor 1024*1024 bytes (guess what - 1,000,000 is exactly what it is - it is NOT 1,048,576)
Look, it's OK to be ignorant. There's no shame in that. I'm sure I am ignorant on any number of topics. However, to cling to a deprecated, failure-prone anachronism -- after being shown the error of your ways -- is downright foolish. And that is indeed shameful.
Stop misusing metric unit prefixes as approximations of binary powers-of-two. Before they kill again.
