Computer Chess Club Archives


Search

Terms

Messages

Subject: Re: UCI versus Winboard

Author: Dann Corbit

Date: 10:54:34 08/22/02

Go up one level in this thread


On August 22, 2002 at 04:59:37, Steve Maughan wrote:

>Dann,
>
>>First off, I would like the protocol manager (e.g. Winboard/Arena/Whatever) to
>>be able to manage/not-manage the opening book,
>
>You can do this using UCI - but to your final point at the end of your email -
>it would be great if we could have a standard opening book editor that could
>produce an opening book useable by any GUI.
>
>>EGTB files, etc.
>
>You can do this using UCI
>
>>I would like it to know if someone claims an illegal move that the move really
>>was or was not illegal.
>
>Doesn't WB and UCI already do this?

Maybe they do, but if so, they make mistakes.  I have seen "dumb" engines make
illegal moves and nobody notices.  I have seen false "illegal move" claims many
times, especially after a lost game; an engine will claim something like e2e4 is
illegal.

>>I would like the protocol manager to be able to manage tournaments.
>
>Shouldn't this surely be the job of the GUI not the protocol?

Depends upon what you expect from the protocol.

>>This is such a common need that I am surprised it is missing.
>>
>>Some of the roles of chess engines are foolishly redundant.  Why must ever
>>program on earth write an interface to the Nalimov tablebase files?  Isn't one
>>good implementation enough if we put it into the protocol manager?
>
>Do you mean the code allowing the engine to interface with the tablebases while
>searching?  UCI already point each engine at the tablebases.  If you do mean the
>search interface code then I'd say this is an engine specific problem.  However,
>it could be improved by better interface documentation and DLLs for the Nalminov
>code.

Engine 'A' wants to swindle.  Engine 'B' has tablebase files too.  At slow time
control, they play through 100 more moves after both parties know for sure it's
a draw.  I absolutely hate that.  And it happens very often.

>
>>The response
>>is:
>>"That's not the role of the protocol manager."
>>My response is:
>>"It should be.  Let's call it the 'game executive' instead."
>
>Doing queries on the Nalminov tablebases while searching is, IMHO, more than
>being a 'game executive'.  It's like asking a tournament arbitrator if he
>wouldn't mind look in his library to see if an endgame is solved!

Exactly.

>>I would also like to see a formal specification for opening books standardized.
>>You can still manage your own book if you like, but it would be nice to have
>>some ready made interface for this tedious task.
>
>Yes I agree.  It would be nice to have an opening book editor that could be used
>by anyone.

I would like everything to be stored in database format so that SQL queries
could be used.

EXECUTE create_tournament "swiss", "5-star"

SELECT EPD_Position, Score FROM book where current_position =
"r1bq1rk1/ppp1bppp/2n1p3/3pPn2/3P4/2P2N2/PP1BNPPP/R2QKB1R w KQ -"
WHERE ELO > 2600
ORDER BY white_success DESC, ELO DESC, score DESC

That sort of thing.



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.