Author: Sune Fischer
Date: 05:13:31 11/15/02
Go up one level in this thread
On November 15, 2002 at 07:23:56, Sune Fischer wrote: I might add, that when I hear "highlevel" I think things like: only one data type. High level means the user doesn't have to worry about whether things are signed unsigned, or floats or integer etc. The data type can handle everything, dividing an integer with an integer will create a float by correctly casting first etc. So if the user really wanted to lose the fraction he would have to call Floor() afterwards. And of course the data type can't overflow, there won't be any stupid 32 bit or 64 bit limitation. You can set the number of significant digits in any computation (like in mathematica and maple), things like that are a given, standard algorithm are there for that. I dread to think about how you would pack things in a hash element? Is it possible to make a compiler that can optimize this as well as when written explicitly C or assembler, personally I would think not. -S.
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.