Computer Chess Club Archives


Search

Terms

Messages

Subject: UCI has several disadvantages

Author: Peter McKenzie

Date: 16:44:15 01/25/02

Go up one level in this thread


I'd just like to point out that UCI isn't clearly better than the winboard
protocol.  Tim Mann has been doing a great job with the winboard protocol
lately, and many probably prefer it to UCI.  I applaud SMK's efforts with UCI,
it does have some nice features for sure, but its not all one way traffic~

Off the top of my head, some advantages of winboard protocol:

a) Hard for the engine to do learning in UCI because you don't know when one
game ends and the next one starts.  Note that the Shredder GUI does its own
learning, the Shredder engine does not.  But it gets difficult if you want your
engine to do its own learning under Shredder.

b) UCI sends the entire move list before each move in the game.  The engine must
play through this list to get to the current position.  This is a bit ugly in my
opinion, and also slightly inefficient.  If you are just starting programming an
engine, its probably a hassle too.

c) Winboard commands can, at a pinch, be typed in manually from a text console.
You can easily play a game like this.  Because of (b) above, this is not
practical in UCI.  Makes manual testing a drag.

d) Winboard has a nice protocol version system, so you don't have to support
everything in the protocol to get up and running.  Can't remember if UCI has
this, don't think so.

e) From memory, pondering in UCI is controlled by the GUI.  The engine can't
switch to pondering another move if it thinks that would be a good idea.

cheers,
Peter



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.