Return-Path: Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id ; Sat, 22 Dec 2001 22:41:19 -0500 Received: (majordomo@vger.kernel.org) by vger.kernel.org id ; Sat, 22 Dec 2001 22:41:09 -0500 Received: from hera.cwi.nl ([192.16.191.8]:38575 "EHLO hera.cwi.nl") by vger.kernel.org with ESMTP id ; Sat, 22 Dec 2001 22:41:02 -0500 From: Andries.Brouwer@cwi.nl Date: Sun, 23 Dec 2001 03:40:56 GMT Message-Id: To: bcrl@redhat.com, cw@f00f.org Subject: Re: Configure.help editorial policy Cc: esr@thyrsus.com, garfield@irving.iisd.sra.com, linux-kernel@vger.kernel.org Sender: linux-kernel-owner@vger.kernel.org X-Mailing-List: linux-kernel@vger.kernel.org Benjamin LaHaise writes: > If you think GB == 1000^3, then please go "correct" all the DRAM > manufacturers out in the world. They just sent me 1GB of ram and > it's coming up as 1073741824 bytes. Oh, please - must we have this same discussion every year? If someone uses an approximate value in some context then that - is desirable if it is more convenient - is admissable either (i) if greater precision is not of interest (or unavailable), or (ii) when the approximate value suffices to derive the exact value uniquely. When things come in power-of-two sizes then no confusion is possible if you use a decimal prefix: in reality the nearest power of two is meant. Thus there is no urgent reason to be precise when talking about the size of memory modules. If you say 1GB, that is an American billion bytes, everybody will assume that in fact the size was 1073741824 bytes. Of course this will change when technology changes and memory comes in arbitrarily sized units, like disks today. These remarks taken together mean that it is not often necessary to use these newfangled abbreviations like GiB or words like gibibyte. Often precision is not required. Or precision is required but the precise size is clear from the context. But it is a serious mistake to doubt G = 1000^3. k, M, G have one and only one meaning. Just like "thousand" has only one meaning and still one can say that this thing, that cost $1049, was bought for a thousand bucks. Disks do not come in power-of-two sizes. So when talking about disk sizes there is no "he says 1000000000 but it must be a power of two so he must mean 1073741824". No, for disk sizes it is just "he says 1000000000 so that is it". In standard texts and other places where there must not be any ambiguity one uses KiB, MiB, GiB. Also in a context where binary and decimal units are both used it is clean to separate them. When Linux boots, I see hdf: 120064896 sectors (61473 MB) w/2048KiB Cache and you see that the MB is decimal and the KiB binary. Good. Concerning this Configure.help stuff, clearly it is not very important what is written, but GiB is slightly better than GB, and I doubt it would lead to any problems. People who have never seen GiB will probably read it as gigabyte, and that is approximately right. Concerning the future of this standard for binary prefixes, it looks like scientists and engineers like them and are careful to distinguish G and Gi; in the open software world programs are slowly adapted to correct usage; Microsoft has not yet heard about the difference between GB and GiB. Andries - To unsubscribe from this list: send the line "unsubscribe linux-kernel" in the body of a message to majordomo@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html Please read the FAQ at http://www.tux.org/lkml/