Author: Fernando Villegas
Date: 08:27:55 05/30/01
Go up one level in this thread
About this I have written more than once some ideas that I called Modular Programming. The idea is that the code must not be runned all at once, but in separated modules activated according ocasion. Let us see your case. If programming is made in just one package of lines, the problem is, as you say, how to ponder different things. Programmers try to solve that with thousand of experimental games and test in order to get an idea of the ideal recipe. But if your program consist in separated modules that comes into action only if some conditions appears, the situation change radically. Somewhere I had a complete explanation of this, I do not know where is in my files now, but for the moment I could say that -this is only an sketchy example- if you have a positional module that is operated only if a tactical one does not deliver an interesting advantage, then your problem is solved at once. Effectively, if the position is such that common rules about bishops does not give a good score anymore, if nothing is decisive or interesting in tactical, middle game kind of game, then a positional module is activated and then and only then consideration as that of squares of same colour of your last bishop appears and are considered. I suppose that this discrimination could be done, if crudely, with just assigning levels of scores to switch from one module to another. By example, if running the convebntional, tactical module the scores are near zero or definitely bad for the program, then automatically he could go to the positional module to see the game in other way. I hope this could be useful for you. my ideas about this are far more complicated, but, as i said before, I do not recall where I left the entire paper. Maybe I erased it because never nobody gave attention to them. Cheers fernando
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.