Author: Miguel A. Ballicora
Date: 08:53:37 08/17/01
Go up one level in this thread
On August 17, 2001 at 09:34:07, Gordon Rattray wrote: >On August 16, 2001 at 16:22:05, Thomas Mayer wrote: > >>Hi Uri, >> >>>>If it's a 32 bit number that is wrapping, this is a problem. The Fritx GUI is >>>>designed to show node counts. Fritz is designed to be able to analyse over long >>>>periods of time. If this doesn't work, it's a bug. >> >>>No >>>A bug is something that the programmer did not mean to do when the code was >>>written. >>> >>>If the programmmer knows about the problem and does not consider fixing the >>>problem as something important then it is not a bug. >>> >>>If the programmer thinks that the program is going to show correct number of >>>nodes then it is a bug. >>> >>>You cannot know if it is a bug or not a bug without asking the programmer. >>> >> >>coolest bug description ever heard... :) I am sure they know the problem >>there... and I am quit sure that it IS a wrapping 32 bit number... some use >>signed variables here, that's why you might see from some engines negativ nps... >>if it is unsigned you have just a lowered nps visible... (The engine anyway acts >>with full speed) - in older days no-one thought that on a micro you can wrap a >>32 bit variabel... not in weeks... :) I know that some have implemented 64 bit >>variabels here now, but it seems to be quit slower - and in chess programming, >>if you have to decide between speed and some eye catching effects it's better to >>tend towards speed... > > >I agree that choosing speed seems like a good design decision, especially for a >chess program. So, if it's too much of an overhead to show a correct node >count, don't show it at all! Or show it for as long as it's valid and then stop >showing it once it is not guaranteed to be correct. Particularly when the latter has to be easy to fix. They might not care to fix the bug, they might not have time to fix the bug or they might believe that it won't increase a penny their profit if they fix the bug, but all those reasons do not mean this is a bug. It is a flaw in the design that causes to show wrong information. If this is not a bug, what is it? Regards, Miguel ed > >Design is indeed about compromises. It is ok to compromise functionality in >order to increase performance. But it is not ok to compromise *correctness* in >order to increase performance. > >Gordon
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.