From: Sam S. <sd...@gn...> - 2012-09-23 15:49:46
|
> * Pascal J. Bourguignon <cwo@vasbezngvzntb.pbz> [2012-09-22 07:32:31 +0200]: > > NEVER USE FLOATING POINT NUMBERS. This is too drastic, IMNSHO. A programmer has to understand them, avoid them when possible, and use them when necessary. E.g., inverting an 1000x1000 integer matrix in rationals will take forever and run out of memory. Doing the same in floats using the standard linear algebra packages (R/Octave/&c) is instantaneous. The basic advice, IMHO, should be "stay in one realm": use only integers or only floats throughout your calculations. -- Sam Steingold (http://sds.podval.org/) on Ubuntu 12.04 (precise) X 11.0.11103000 http://www.childpsy.net/ http://jihadwatch.org http://palestinefacts.org http://www.PetitionOnline.com/tap12009/ http://openvotingconsortium.org Computers are like air conditioners: they don't work with open windows! |