Computer Chess Club Archives


Search

Terms

Messages

Subject: Re: Problems with the EPD "ce" opcode

Author: Steven Edwards

Date: 20:07:52 04/25/05

Go up one level in this thread


On April 25, 2005 at 16:15:23, Dieter Buerssner wrote:

>Steven, I agree with your points about the deficiency of "ce". Personally, I
>like the UCI method. For the score, besides the typical pawn value evaluation,
>it spells out mate scores, and also can flag the score as upper bound or lower
>bound. Did you have a look at the UCI protocol? Perhaps, that scoring might fit
>naturally to EPD. Perhaps even giving two bounds might be useful (for some
>engines). Saying something like a <= score <= b. FOR Things like illegal,
>perhaps another tag might be useful.

I looked at the UCI spec a while ago and I agree with Hyatt that such an
interface attempts too much control over a client program and also assumes too
much about a client's internal organization.  Also, I seem to remember that it
used the old coordinate notation for moves instead of SAN.

Let's not ask for too much of what should be a simple interface standard whose
external representation should be easily read by humans.



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.