"We already often informally use 'k' (kilo) for 1000, and, ... clearly-defined meanings and so there is no risk of confusion."
"Well, the "informal" use of K for 1000 very neatly demonstrates the problem with your proposal."
Ah, but I didn't write 'K', I wrote 'k', and even clearly specified that I meant 'kilo', which is clearly defined as 10^3, as in kilogram.
Indeed, there does still exist some confusion caused when 'K' (note case) came into use to mean "1000-ish" (specifically, 1024, or 2^10, as you know) in binary computing terms. When computer memory and storage were still mostly counted in "kilobytes" (sic) the difference between 1000s and 1024s obviously wasn't deemed to be enormously significant, but now that orders of magnitude further up the scale are much more commonplace, the factor of error becomes considerable.
To avoid this confusion, new binary prefixes have been invented for use with binary quantities (such as computer memory), and the erroneous use of the standard decimal prefixes for this purpose is now discouraged (although adoption is being somewhat slow!).
Hence, for 1 KB (of memory) you should instead write 1 KiB (kibibyte), similarly, MiB/mebibyte, GiB/gibibyte, TiB/tebibyte, and so on. (The prefixes are formed by condensing the first part of kilo, mega, etc, with the first part of 'binary'.)
Note, however, that for whatever bizarre reason (1), hard disk manufacturers decided that when they wrote MB, GB, etc, they *did* mean decimal scale factors (multiples of 1000) and not binary ones (multiples of 1024), so MB, GB, etc, should still correctly be used there, rather than MiB, GiB, etc.
(1) probably because it meant they could use larger numbers and so make their disks seem larger than they really were.
"If I applied for, and got, a job which was advertised at a salary of £30K, I would be somewhat ... only £30,000 instead of £30,720! (Especially since any job that I applied for would be in the world of computing.)"
But why would you be applying binary scales to monetary values, which are decimal? I think if you applied that logic, you wouldn't actually get the job ;-)
(A memory manufacturer might like your rationale (but I still doubt they'd pay you the difference), but a hard disk manufacturer (kB) or a network connectivity supplier (kbit), or a CPU company (kHz) definitely wouldn't be impressed!)
If we were being pedantic (and of course we are, because this is ucle) then surmising that 'K' = binary-kilo, because it is not written 'k' = decimal-kilo, would also indicate an error on the part of the writer, because 'K' (when it does not mean kelvin, which it clearly cannot here) should really be 'Ki' were it to mean what is now called 'kibi-'.
..which is why I expressly wrote 'k' in the first place! :-)
"If we follow your suggestion and simply redefine K as 1,000, M as 1,000,000 etc., it does rather echo the urban legend that Alabama legislated to change the value of pi to 3."
Hmmph. ;-)
But 'k' *does* mean 1000, and 'Ki' *does* mean 1024, so there's no redefining going on.. :-)
David.
David M. Edinburgh, Scotland. (en,fr,(de)