Relay-Version: version B 2.10 5/3/83; site utzoo.UUCP Path: utzoo!mnetor!seismo!cmcl2!phri!roy From: roy@phri.UUCP (Roy Smith) Newsgroups: net.arch Subject: Re: VERY LARGE main memories: crypt Message-ID: <2435@phri.UUCP> Date: Fri, 12-Sep-86 16:01:00 EDT Article-I.D.: phri.2435 Posted: Fri Sep 12 16:01:00 1986 Date-Received: Sat, 13-Sep-86 21:25:03 EDT References: <1178@bu-cs.bu-cs.BU.EDU> <1087@ihwpt.UUCP> <553@astroatc.UUCP> <2431@phri.UUCP> <668@hropus.UUCP> Reply-To: roy@phri.UUCP (Roy Smith) Organization: Public Health Research Inst. (NY, NY) Lines: 32 Summary: apology for erroneous electron statistics. I claimed that 200 bits is a good upper limit on address size because there are 2^200 electrons in the universe. I was taken to task on this and it was suggested that I meant 10^200. I reiterated that I meant 2^200, which is what other people would call about 10^40th. In article <668@hropus.UUCP> jin@hropus.UUCP (Jerry Natowitz) pointed out that 10^40 is off by rather a lot. My apologies. I used a calculator program on Unix to calculate 2^200 and came up with 1.7e38. Stupid me, by now I should recognize that number as something magic -- the largest number you can represent in floating point on a Vax. It seems that 2^200 produces floating point overflow and the calculator program I used doesn't catch overflows. And yes, I did think 1.7e38 was a bit low (I was expecting more like 1e60, or about 3.something bits per decimal digit). There is a discussion going on in RISKS-DIGEST (aka mod.risks) about people trusting computers too much. Now I find myself just as guilty. Mea Culpa. Anyway, I stand by my assertion that there are 2^200th electrons in the universe. If you prefer to work in decimal, work out the conversion yourself. For the purposes of this discussion, I consider any arithmetic error in that conversion (even a factor of 10^20!) to be trivial. A reasonable argument that 200 bits is not a good upper bound has been made by various other people, however. There are often times when a sparse address space is useful. If you take my home address as a number, RoySmith222UnionStreetBrooklynNY11231, you need more than 200 bits (7-bit ASCII). For a world data base, it might be convienent to use this directly as an index into a table, and the hell with hashing. -- Roy Smith, {allegra,philabs}!phri!roy System Administrator, Public Health Research Institute 455 First Avenue, New York, NY 10016