Date: Thu, 21 Feb 2002 20:34:13 +0000 (GMT) From: Andrew McKay <andy@openirc.co.uk> To: doc@freebsd.org Cc: Michael Wardle <michael.wardle@adacel.com> Subject: Re: inconsistent use of data units Message-ID: <20020221200735.S76448-100000@fluoxetine.lan> In-Reply-To: <1014297857.177.20.camel@cocaine>
next in thread | previous in thread | raw e-mail | index | archive | help
I agree that the original premise is correct and that using SI prefixes carries a certain suggestion that they operate in the same way that SI prefixes usually do. However. Computer science is a scientific field which is unique in that it operates in powers of two rather than powers of ten. This is due to the nature of binary mathematics, for which powers of ten are simply illogical. A byte does NOT consist of ten bits and never will. (none of this is intended to be news to anyone - it is a mere statement of fact) SI units define k, M, G &c to refer to increasing 10^3 multiples within decimal systems. Computer scientists define K, M, G &c to refer to increasing 2^10 multiples within binary systems. What works for apples does not apply equally to oranges and neither should it be made to try. Given the prevailing conditions within the field of computer science it would, I believe, be a fruitless endeavour to attempt to merge a long standing de facto standard with a long standing ratified standard. If push came to shove I suspect that the overwhelming majority of computer scientists would rather form their own standards body and 'ratify' the use of K, M, G &c to refer to increasing 2^10 multiples than convert to a system whereby they refer to increasing 10^3 multiples. The reasons for this are clear. Unless you are selling something and wish to make your product appear larger than it is then you have ABSOLUTELY no need to refer to 1000, 1000000 or 1000000000 bytes. They are utterly useless. 1024 and its multiples, however, are exceedingly useful. Furthermore, given the vast percentage of 'computer people' (scientists and users alike) who understand the system as is compared with the tiny percentage who know or even care that it is not 100% technically correct, any change that was decided to be made would meet such resistance and take so long to implement that the benefits would not be worthwhile. Basically, the system is not broken. The only breakage occurs because it deviates from a system which does not apply to it, from which it borrowed handy prefixes as an approximation for their ultimate use. The people who do not understand this difference, who would be likely to care, are in an infinitessimal minority. So, imho, the only question to answer is 'who is going to go through the doc tree and apply a standard, consistent with that which is understood by the majority of people, to the docs?' I would offer but I already have a large number of jobs on my list :) Andy -- Andrew McKay <andy@openirc.co.uk> To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-doc" in the body of the message
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?20020221200735.S76448-100000>