Author: Sune Fischer
Date: 14:24:41 04/09/03
Go up one level in this thread
On April 09, 2003 at 17:04:33, Russell Reagan wrote: >On April 09, 2003 at 15:38:13, Jonas Cohonas wrote: > >>That must be "UCI" you are referring to, personally i think it gives the average >>user more possibilities, not to mention ease of use. For the more experienced >>user winboard might be better, i don't know though. > >IMO, this debate is along the same lines as actors vs. directors vs. screen >writers, etc. The screen writer wrote his script, and it's that person's baby. >They don't want some director or producer adding in fluff or taking things away. >The director usually has a big ego, so they want to put their mark on a film, >impose their "style" on it. The actor also has a big ego and thinks they know >best how to act a part, and they also want to leave their mark on the film. Each >person sees it from a different point of view and each has different (usually >self serving) motives. The Winboard vs. UCI argument is similar, I think. > >The user wants features and wants to be able to configure anything and >everything to their liking, without having to learn how to program, so things >like specifying ponder off, hash size, etc. via the GUI is something a user >wants. UCI is probably the closest thing to this (aside from commercial >protocols maybe). > >The GUI/protocol creator wants to add support for all kinds of features, because >this person is usually more "user centered". > >Then you have the engine programmer, who has his own ideas about what he wants >to user to be messing with. Most engine authors aren't concerned with what the >user wants. They are more concerned with their engine being as strong as >possible, and so they don't want the GUI/user/protocol to interfere or limit >what they can or can't do (now or in the future). > >I think most engine authors view GUI's as simple methods to play engine-engine >matches, or simple ways to play against people on ICS's. If they want to change >a feature or configure something, they probably change a config file or >recompile. > >Personally, I think the GUI shouldn't tell your program what to do. I think it >should be the other way around. The engine should be able to tell the program >what to do, how to look (colors, etc.). Maybe the engine author would like to >see the current coverage of the board by one side (for developing a new >evaluation parameter, or whatever), and if the GUI was really just that (a >graphical user interface to the ENGINE), then the engine could highlight certain >squares or whatever it chose. > >I guess it all boils down to how you see things. Is the GUI serving the engine, >or is the engine serving the GUI? Well said, I think you are spot on. Though, I see no problem in taking all the cool stuff from UCI and add support for it in WBIII. The special engine "options" (e.g. evaluation parameters so we don't need to go to config files), well defined mate scores, signs on the score, format of the PV, path for the end game tables and size of hash, naturally all that should be configured from within the GUI. Of course we keep the force and go mode, keep sending "result" to aid the learning and allow the engine to ponder the way it wants to. I see no reason why WBIII couldn't be the best of both worlds. -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.