Author: Heiner Marxen
Date: 10:44:39 05/16/01
Go up one level in this thread
On May 15, 2001 at 22:43:45, Dann Corbit wrote: >On May 15, 2001 at 22:09:28, leonid wrote: Hello Leonid, hello Dann! >>To be sincere, Heiner, I even more impressed by your solver. It is only a pitty >>that it is hidden of its clumsy shell. Without any doubt the best mate solver >>ever! I hope that some programmer will come with graphics and very soon. This >>will make it instantly accessible and popular with everybody. The most difficult >>and laborious part of mate solver is already accomplished. And we have here so >>many people that are very prolific in C and graphics! I hope somebody is >>listening... Well, _I_ am listening. As far as I can tell from the feedback regarding Chest (since I put it on the web), there are very few people who try to hack the sources. I know of only one case since Dec'99. I would be pleased to be proven wrong, but I expect I will have to do it myself. Ok, I understand that there are some people who would like a GUI for Chest. I also understand, that an existing GUI would possible attract many more to use Chest. I want to explain my situation, and will speculate about future directions. Chest is no product in any sense. I do not receive anything (except some sort of "fame", perhaps) for any work on it. I have so far worked on Chest only for my own fun. _That_ is what I get back. My own interest is not with the interface of Chest, but with the inner workings/methods/algorithms. Of course, it has to have an interface. Since I'm for 20 years a professional programmer working with all sorts of UNIX (BSD, HP-UX, IRIX, Solaris, Linux and many more), that is the (software) world I live in. I never used a Windows system. And since I hate MS (no, please lets not discuss this point), I have no plans to walk into the Windows world. I have nontheless tried to make the existing Chest workable under Windows. Dann has helped me. In fact, he ported it, and I have integrated his changes. Still, it is just a "console application". Using Chest with any UNIX is not any different. "console applications" is the type of tool, which is common with UNIX. I do that all day. So, I'm happy with it. Once we had a GUI for Chest. But that was quite special, and would not help any "user", it was helpful to the development team, to noone else. Also, it was not very portable even among UNIX flavours, and on those systems I use today, it does not compile any more. Currently, I use it in this text based way. When Leonid posts some mate, I start an editor to create yet another text file, containing input for Chest. Here is the last example, the file "i.ccc168": > Posted by leonid on May 15, 2001 at 06:30:34: > Subject: One mate to solve... > http://www.icdchess.com/forums/1/message.shtml?169798 LE F 1qbBBbq1/1Q5r/R2Nq1P1/2Q1N1n1/KQ2q1rk/2Q1N1n1/R2Nq1P1/1Q5r w - - z9w .. Most of the above is entered with copy/paste from the browser window to the editor window. Then I call Chest and let it append to that file: ./chest -sSLU i.ccc168 >>& i.ccc168 & Now it is working in the background. The last lines tells me about the depth already reached, e.g. I might see 2 minutes later at the end: # 5 2.46 1.18 11330- 0 # 6 17.65 1.34 86730- 0 # 7 100.17 1.56 504249- 0 That lets me get an idea how long it will take to complete depth 9. You see, I do not need anything more. Ok, lets assume a GUI is badly needed. Lets demand, that it must run under Windows, where most of the "users" are. Lets further assume, I should implement that. What does that imply? First, I will not create a complete GUI from the start. Reasons are many: - As a matter of fact, I have never programmed a GUI. Yes. Sorry. - Even if I could do it, I would not like to. I would rather try to interface with an existing GUI. That is the way xboard/winboard works. It is the right way to do it. - It demands too much time from me, for something I would not even use myself. - If the GUI would fit my own needs as a developer, it most probably would not please any "user". But that is our current goal. Then, it must be portable. I.e. I will demand that it also runs with Linux/UNIX and not just with Windows. That does not leave many options. My only chance seems to be (as far as I know, teach me something else), to interface to some existing GUI. Ok, I have just downloaded the newest version of xboard/winboard. As far as discussed above, it fulfilles all demands. But... Xboard is an interface to play chess. I'm not sure, whether it can be sensibly used to interface a mate solver. Remember? Chest does not play a game of chess. I can imagine to get a board from the GUI, a color to move, even a search depth, and then get a command to think about a best move. Fine so far. When I find a mate, I report the key move as best move, and also report a PV. Ok. Now, what shall Chest report to the GUI, when there is no mate found? In any case, within the xboard package I have found a long description of the engine interface/protocol. I promise to read that carefully, in order to find a meaningful way to interface to chest. Any hints and suggestions are welcome. Are there any other chances to interface Chest to a GUI? >I like a simple interface with a lot of options. Anyway, you could easily >interface Chest to Chess Assistant and have the most glorious GUI interface >imagineable. Indeed, the EPD processor type engines can bolt right in to Chess >Assistant. Chest is a very good example of a chess analyzer that eats EPD and >spits out EPD. Could you elaborate on that, Dann? What is the engine interface of CA? Could I implement that? Is it already there, and we only need deteailed instruction how to set it up? Wher can I read about it? >I share your respect for Heiner's coding. He's also a real nice guy. Thanks! I try to live up to that description. But I need some help from others, who know better about GUIs. I myself know next to nothing about GUI programming and interfacing. When there is a way I can do it with moderate efforts, such that I can run it under Linux/UNIX myself, I will seriously consider to go that way. Excuse the length of the above... I felt it necessary to explain myself. <eagerly waiting for many helpful responses> Regards, Heiner
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.