Author: Reinhard Scharnagl
Date: 12:41:18 01/06/04
Go up one level in this thread
On January 06, 2004 at 15:17:42, Russell Reagan wrote: >On January 06, 2004 at 13:48:17, Reinhard Scharnagl wrote: > >>But I see a need for a more >>modern protocol supporting also 10x8 boards, exotic chess pieces and a unic >>encoding of moves, not like Winboard is doing it with castlings. > >Encoding things in a binary form is not a very good choice for a communication >protocol in this situation. Plain text is a better choice if your protocol is >going to be an open protocol that is going to be used by many people. Using >things like DLLs, COM, etc. do not make your protocol better just because they >may be more "modern" than using plain text pipes. In fact, I think those things >make it worse, because all they do is restrict the protocol to a single >operating system, and you also assume that the engine will be written in a >handful of languages. If you look at what communication protocols are successful >(in any field, not just computer chess), they don't restrict program writers to >a single operating system or programming language. Text pipes do a very good job >of this. If an operating system and programming language can write to standard >output and read from standard input, then it can support the Winboard or UCI >protocols. A human could even interact with the program by typing commands into >it (you can send an email using a telnet client, because the Simple Mail >Transfer Protocol, SMTP, is plain text). > >Of course, if you are only using this protocol for your own personal use, then >you can make it as restrictive as you want, but don't be suprised when not a >single other person in the world uses it. If you want other people to use it, >then make it easy for them to use. That probably means using plain text pipe >communication. > >What is it that you don't like about the way Winboard handles chess moves? I >think it should be very easy to extend the Winboard protocol to handle boards of >different sizes and new pieces (to handle something like Gothic chess). Talk to >Tim Mann about it and see what he thinks. He would probably welcome your ideas >and any help you can offer. One year ago I tried to start a discussion on such themes (may be in the Winboard forum, I don't know today). But I had to learn, that trying to discuss anything concerning Winboard status quo was understood as a sort of aggressive military act. So I lost the hope for something to be changable within Winboard. E.g. I do not see any need for distinguishing between so called variants NORMAL, NOCASTLE and FISCHERANDOM. All important data could be expressed by an intializing FEN string. And there is no need for to encode castlings in a different way, when associated with FRC. A good FRC implementation always will be playing FRC, even when starting from a classic chess starting position. So there is absolutely no need to complicate a protocol with such artificial distinctions. Regards, 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.