Author: Dann Corbit
Date: 16:08:55 01/25/02
Go up one level in this thread
On January 25, 2002 at 18:52:57, Mike Hood wrote: >On January 25, 2002 at 17:49:11, Dann Corbit wrote: > >>On January 25, 2002 at 17:37:15, Mike Hood wrote: >> >>>From all that I've read about UCI, it seems to have several technical advantages >>>over Winboard. >> >>Name one. >> >>>But, as we all know, the best product doesn't always succeed. >>> >>>There will only be large support among chess engine programmers if there is a >>>high quality free GUI available. (If UCI interpreatation remains proprietary to >>>Chessbase it will kill the protocol before it even takes off). Is there a >>>comparable freeware equivalent of the Winboard program (WinUCI)? If not, are >>>there any plans to create one? >> >>Why not trash UCI and all the other pretender interfaces, and go with the open, >>well adopted, flexible specification: >>Winboard. > >Thanks for your thoughts, Dann, but you didn't really answer my question. This >is my point: > >Winboard has one big advantage over UCI. If a programmer wants to write a new >Winboard engine from scratch he downloads the attractive Winboard GUI for free >and gets stuck into the programming. Tim Mann's software is well tested, and so >the programmer can put all his effort into debugging his own code. > >But what about UCI? If a programmer wants to write a UCI engine, his first step >is to spend money on Shredder or another Chessbase product. This is enough of a >disadvantage to encourage the programmer to stay with Winboard. If the Universal >Chess Interface is to be promoted seriously, there MUST be a freeware GUI made >available. My point is: "Why do we want to promote a different standard?" It's already a lot of bother to write a Winboard interface for a program. In fact, getting that part just right is nearly as hard as writing the chess engine. Now, we add another protocol? Worse than a waste of time. I think they should instead add any useful missing features to the Winboard protocol. Or (if there is to be another standard) there should be a truly compelling reason for adopting it. One thing that might make the UCI attractive is if someone wrote a complete and accurate reusable interface for it. Then, all your program would have to do is hook up to the functional API that is already written. Even at that, I don't like having many different languages that the chess engine must learn to speak. It would be much better if they could all speak the same language. If there is anything missing from the Winboard protocol, then add it as an extension. If it is worthwhile, eventually it would be adopted. It's not exactly a tower of Babel, with just a few specs for chess interface engines. But imagine if there were two C language standards, or if ISO and ANSI used different ones. How will I write portable, correct code in that instance. In the same way, I would vastly prefer a single, coherent standard for chess engine interface. Winboard is open. Winboard is well documented. Winboard is widely implemented. Winboard can be adopted as an excellent interface into professional database products. In fact, Chess Assistant has done this. So UCI is completely unnecessary.
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.