Author: Reinhard Scharnagl
Date: 02:44:09 03/11/05
Go up one level in this thread
On March 11, 2005 at 05:08:22, Volker Annuss wrote: >>I prosume that there are equivalent constructs to Windows DLLs in other OSs. >>The concept of dynamically binding modules could hardly be a secret. Hi Volker, >Keep it simple and stupid. I regard the handling of IO streams synchron to computing not to be small and simple. >With a protocol that uses console input and output you don't have to develop >different versions of an engine when you want to support more than one protocol. Before that would be a problem for an engine supporting FRC and 10x8 there first some GUIs have to exist. But the lack of such GUIs forced me to start a solution by myself. >And by using console input and output you can do some testing without a GUI. This argument is valid. But if your engine is working, this plus has vanished. Using IO could be a solution if the existence of threads would be assumed. Personally I would welcome any consistent protocol supporting engines declaring themself FRC aware (by defining an option "FRC" as possible in UCI) or 10x8 aware (by defining an option "10x8", meaning also supporting 10x8 and Capablanca extended piece set). Reinhard.
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.