Author: Dieter Buerssner
Date: 12:24:47 04/25/05
Go up one level in this thread
On April 21, 2005 at 18:39:02, Dann Corbit wrote: >http://www.americanscientist.org/content/AMSCI/AMSCI/ArticleAltFormat/20035214317_146.pdf Thanks for the link, Dann. I did not have time yet, to read it very carefully, but I looked over it. It seems to boil down to the method, how to define storage needs. "r*w" is given in the paper. I am not sure yet, whether I agree, that this is a sensible definition. Say, I want to store the number 255. In base 2, this gives 8*2=16. In base 16 this gives 2*16=32, clearly worse. But we know already, that a base 16 digit can be stored with a device, that only needs 4 base 2 digits. So, assuming a multiplier of 16 here does not look right. Or assume another device. Say a hydrogen atom. The ground state shall be 0, excited state shall be 1. (Of course, it will not be easy, to prepare a hydrogen atom, that stays in excited state ...). We will need this one hydrogen atom to store a binary digit. But there are many excited states. If we could prepare and distinguish several excited states, the space would not change. Or look at written language. Sure this is is a bit far fetched ... We could note binary numbers and use characters "0" and "1". But with the same space we could use all upper and lower case letters, and also all digits, and many more glyphs, that are easily distiguishable. So we easily get base 256 or so (Chinise people would estimate a much larger base) with practically the same space. Making the space needed proportional to the base would be like saying: Write the (decimal) digit nine 111111111 and the decimal digit one as ........1 If we really design the storage device in this way, base three will be most efficient. To me, this assumptions looks rather arbitrarily, and away from practice. Regards, Dieter
This page took 0 seconds to execute
Last modified: Thu, 15 Apr 21 08:11:13 -0700
Current Computer Chess Club Forums at Talkchess. This site by Sean Mintz.