CDRLabs.com

dodecahedron wrote:since a major part of the computing community thinks of 1K=1024, the hard drive makers are misleading IMO.

true, it might be better if everyone moves over to standard SI units, 1K=1000 etc. but until this move had been made these are deceptions.

and you can't really think that the drive makers are doing what they're doing in order to promote a move to SI. such a move will have to originate from a more authoritative and acceptable company/organization.

Oh yes, of course the manufacturers aren't doing this to promote the SI standards. They're merely using the standards to benefit themselves. But isn't that what business is all about? There is no deception here.

It is true that the majority of the industry follows a de-facto standard, which started out as a hijacking of the SI standards in the first place. The word "kilo" means thousand. It is taken from the Greek word "khlioi" and has been used by the SIU for a very long time. The fact that a bunch of geeks in the 1960s decided to corrupt its meaning and that has now been accepted by people does not make it a standard.

But I have to make an exception in this case: Imperial measures are arbitrary (based on the size of the king's foot, the king's favourite pot) etc, while binary measures are a far more accurate representation of data storage than SI measures are. (Assuming SI even is too stupid to recognise the binary standard.)

Although, jase has another point about the word "kilo" and its meaning. By definition, a kilobyte should be 1000 bytes.

jj_frap wrote:But I have to make an exception in this case: Imperial measures are arbitrary (based on the size of the king's foot, the king's favourite pot) etc, while binary measures are a far more accurate representation of data storage than SI measures are. (Assuming SI even is too stupid to recognise the binary standard.)

All measurements are arbitrary. The SI units tend to be more sensible than rival formats, (although it can be argued that the imperial measures are easier to work with in some instances as they often divide by 2, 4 etc more easily than metric) and the measures of mass, volume, length etc tend to work together more fluently, but they are just as arbitrary, with one or two exceptions such as temperature, as any other measurements.

But I don't agree that a measurement using base-2 maths is any more accurate than one using base-10. In both cases you are dealing with empirical data; the standards may be easier to work with than the SI measures but they are no more or less accurate; the "kilobyte" is arbitrary in any case -- why not use 16 or 256 bytes as your next point of reference rather than 1024, as that would be more sensible if you think about it? (16=10 hex or 1 0000 binary; 256=100 hex or 1 0000 0000 binary whereas 1024=400 hex or 100 0000 0000 binary).

It's something I've thought since programming on the old 8-bit machines. A kilobyte is such a stupid measurement in any case, it means nothing in assembly. The numbers which matter, of course, are

1, 16, 256, 4096, 65536, 1048576 (=a gibibyte funnily enough) and so on.