Spider Chess again ...

[Frank Quisinsky]

Geschrieben von: / Posted by: Frank Quisinsky at 08 November 2001 01:39:36:
Hi,
the very sympathical programmer from Spider Chess, Martin Giepmans (Netherland) wrote me:
"SpiderChess has it's own interface and the program is not compatible with WinBoard. But there will probably be a Winboard-version somewhere in the future.
The coming months I do not have much time for work on SpiderChess, so, well,
maybe next year. I hope so."
It is possible that he public his own GUI also in the future ...
Good news, Spider Chess is very strong, I believe over 2.300 ELO (think so after the games which I saw). His own GUI is beautiful for me ...
Best
Frank

[WYx]

Geschrieben von: / Posted by: WYx at 08 November 2001 09:17:03:
Als Antwort auf: / As an answer to: Spider Chess again ... geschrieben von: / posted by: Frank Quisinsky at 08 November 2001 01:39:36:

Hi,
the very sympathical programmer from Spider Chess, Martin Giepmans (Netherland) wrote me:
"SpiderChess has it's own interface and the program is not compatible with WinBoard. But there will probably be a Winboard-version somewhere in the future.
The coming months I do not have much time for work on SpiderChess, so, well,
maybe next year. I hope so."
It is possible that he public his own GUI also in the future ...
Good news, Spider Chess is very strong, I believe over 2.300 ELO (think so after the games which I saw). His own GUI is beautiful for me ...

SpiderChess is Borland Pascal engine, AFAIK.
With this language is very strong to implement WB protocol.

I think, just 2100 !
WYx

[Tony Werten]

Geschrieben von: / Posted by: Tony Werten at 09 November 2001 10:53:19:
Als Antwort auf: / As an answer to: Re: Spider Chess again ... geschrieben von: / posted by: WYx at 08 November 2001 09:17:03:

Hi,
the very sympathical programmer from Spider Chess, Martin Giepmans (Netherland) wrote me:
"SpiderChess has it's own interface and the program is not compatible with WinBoard. But there will probably be a Winboard-version somewhere in the future.
The coming months I do not have much time for work on SpiderChess, so, well,
maybe next year. I hope so."
It is possible that he public his own GUI also in the future ...
Good news, Spider Chess is very strong, I believe over 2.300 ELO (think so after the games which I saw). His own GUI is beautiful for me ...

SpiderChess is Borland Pascal engine, AFAIK.
With this language is very strong to implement WB protocol.

I think, just 2100 !
WYx

Delphi to be precise.
Do you mean it's difficult ? Not really, at Tims site there is some samplecode. Basic winboard functionality can be done in 30 minutes.
Tony

[WYx]

Geschrieben von: / Posted by: WYx at 09 November 2001 11:59:45:
Als Antwort auf: / As an answer to: Re: Spider Chess again ... geschrieben von: / posted by: Tony Werten at 09 November 2001 10:53:19:
Hi Tony

SpiderChess is Borland Pascal engine, AFAIK.
With this language is very strong to implement WB protocol.

Delphi to be precise.
Do you mean it's difficult ? Not really, at Tims site there is some samplecode.
Basic winboard functionality can be done in 30 minutes.

In Leiden 2000 was Borland Pascal. Since then I don't know.

for Delphi but not for Borland Pascal.
Has a difference.

Yes?
You are in practice in development under WB, but maybe the Author of SpiderChess (SpiderGirl) not.
For me was that 2 days.

WYx

[Lyapko George]

Geschrieben von: / Posted by: Lyapko George at 13 November 2001 12:07:09:
Als Antwort auf: / As an answer to: Re: Spider Chess again ... geschrieben von: / posted by: WYx at 09 November 2001 11:59:45:

With this language is very strong to implement WB protocol.

Do you mean it's difficult ? Not really, at Tims site there is some samplecode.
Basic winboard functionality can be done in 30 minutes.

for Delphi but not for Borland Pascal.
Has a difference.

Yes?
You are in practice in development under WB, but maybe the Author of SpiderChess (SpiderGirl) not.
For me was that 2 days.

WYx

The first version of Bestia was written in TP6 and to make it basically WB compatible(time,otim,white,black,new,go,force,move,result etc.) it tooks me about 10 minutes (KeyPressed and Readln). When I switched to Delphi it was far more complicated...
Best regards,
George

[WYx]

Geschrieben von: / Posted by: WYx at 13 November 2001 13:01:52:
Als Antwort auf: / As an answer to: Re: Spider Chess again ... geschrieben von: / posted by: Lyapko George at 13 November 2001 12:07:09:

With this language is very strong to implement WB protocol.

Do you mean it's difficult ? Not really, at Tims site there is some samplecode.
Basic winboard functionality can be done in 30 minutes.

for Delphi but not for Borland Pascal.
Has a difference.

Yes?
You are in practice in development under WB, but maybe the Author of SpiderChess (SpiderGirl) not.
For me was that 2 days.

WYx

The first version of Bestia was written in TP6 and to make it basically WB compatible(time,otim,white,black,new,go,force,move,result etc.) it tooks me about 10 minutes (KeyPressed and Readln).

Yes?
I know so, TP6 was't compiling win32 code.
And WB protocol works with win32 EXE only.

When I switched to Delphi it was far more complicated...

Yes, sure.
WYx

[Lyapko George]

Geschrieben von: / Posted by: Lyapko George at 14 November 2001 11:26:53:
Als Antwort auf: / As an answer to: Re: Spider Chess again ... geschrieben von: / posted by: WYx at 13 November 2001 13:01:52:

I know so, TP6 was't compiling win32 code.
And WB protocol works with win32 EXE only.

This is false statement.
WinBoard works with ANY DOS executable.
Bestia v0.0 was only 27kB large, used no hash and was about 1900 ELO only.
The whole source code is 55 kB in pure Turbo Pascal (it had no ASM statements).
Just for fun I ran a little bullet match(2min/40 moves) under Winboard 4.0.7:
Bestia v0.0-Crux v0.50i +1-3=0
Best regards,
George

[WYx]

Geschrieben von: / Posted by: WYx at 14 November 2001 11:54:43:
Als Antwort auf: / As an answer to: Re: Spider Chess again ... geschrieben von: / posted by: Lyapko George at 14 November 2001 11:26:53:

I know so, TP6 was't compiling win32 code.
And WB protocol works with win32 EXE only.

This is false statement.
WinBoard works with ANY DOS executable.
Bestia v0.0 was only 27kB large, used no hash and was about 1900 ELO only.
The whole source code is 55 kB in pure Turbo Pascal (it had no ASM statements).
Just for fun I ran a little bullet match(2min/40 moves) under Winboard 4.0.7:
Bestia v0.0-Crux v0.50i +1-3=0

Cutting from WB documentation:
5. WinBoard requires Win32 engines
Due to some Microsoft brain damage that I don't understand, WinBoard does not work with chess engines that were compiled to use a DOS extender for 32-bit addressing. (Probably not with 16-bit DOS or Windows programs either.) WinBoard works only with engines that are compiled for the Win32 API. You can get a free compiler that targets the Win32 API from . I think DJGPP 2.x should also work if you use the RSXNTDJ extension, but I haven't tried it. Of course, Microsoft Visual C++ will work. Most likely the other commercial products that support Win32 will work too (Borland, etc.), but I have not tried them.
I don't know, works or not with tp6 generated EXE, I have not tried!
I recline upon the docuemntation just.


Please, send me in email this tp6 version, and I'll review.
WYx

[Dieter Buerssner]

Geschrieben von: / Posted by: Dieter Buerssner at 14 November 2001 12:49:07:
Als Antwort auf: / As an answer to: from WB documantation... geschrieben von: / posted by: WYx at 14 November 2001 11:54:43:

This is false statement.
WinBoard works with ANY DOS executable.

Cutting from WB documentation:
5. WinBoard requires Win32 engines
Due to some Microsoft brain damage that I don't understand, WinBoard does not work with chess engines that were compiled to use a DOS extender for 32-bit addressing. (Probably not with 16-bit DOS or Windows programs either.) WinBoard works only with engines that are compiled for the Win32 API. You can get a free compiler that targets the Win32 API from . I think DJGPP 2.x should also work if you use the RSXNTDJ extension, but I haven't tried it.

I think the WB documentation is not totally correct. For me a DJGPP 2.x compile works (sort of) even without RSXNTDJ. At least a normal game without pondering will work.
It was already discussed here recently, whether it is possible to check for input in the pipe or not. I didn't try. But I know, that DJGPP comes with a select() function, and perhaps there is some hope, that this works for the pipe between WB and the engine. Another possibility might be signals, but I did not check. Signals are to some extend administrated by the DPMI (DOS protected mode interface) of Windows, and it seems not impossible that a Win executable can send a signal to a DOS-extender executable.
Regards,
Dieter

[Miguel A. Ballicora]

Geschrieben von: / Posted by: Miguel A. Ballicora at 14 November 2001 17:16:15:
Als Antwort auf: / As an answer to: Re: from WB documantation... geschrieben von: / posted by: Dieter Buerssner at 14 November 2001 12:49:07:

This is false statement.
WinBoard works with ANY DOS executable.

Cutting from WB documentation:
5. WinBoard requires Win32 engines
Due to some Microsoft brain damage that I don't understand, WinBoard does not work with chess engines that were compiled to use a DOS extender for 32-bit addressing. (Probably not with 16-bit DOS or Windows programs either.) WinBoard works only with engines that are compiled for the Win32 API. You can get a free compiler that targets the Win32 API from . I think DJGPP 2.x should also work if you use the RSXNTDJ extension, but I haven't tried it.

I think the WB documentation is not totally correct. For me a DJGPP 2.x compile works (sort of) even without RSXNTDJ. At least a normal game without pondering will work.
It was already discussed here recently, whether it is possible to check for input in the pipe or not. I didn't try. But I know, that DJGPP comes with a select() function, and perhaps there is some hope, that this works for the pipe
Regards,
Dieter

A djgpp compile worked perfect for me in my machine and it was 25% faster
than a cygwin compile. However, it did not work in other machines for reasons
that I do not know. That version was 100% ANSI-C though (so, no ponder of any kind).
I think that the bottom line is that a DOS program can work but it is not guaranteed, even if it does in one machine. All different flavors of Windows
might make things more difficult to predict.
IIRC, that select() function is a "dummy" function. It does not do anything and it is present so programs that are being ported can be compiled for diagnosis
without firing compiler errors. But I might be wrong. I read that somewhere, I think in the help provided with RHIDE (the free interface).
All this is a pitty because I liked djgpp a lot.
Regards,
Miguel

between WB and the engine. Another possibility might be signals, but I did not check. Signals are to some extend administrated by the DPMI (DOS protected mode interface) of Windows, and it seems not impossible that a Win executable can send a signal to a DOS-extender executable.

[Dieter Buerssner]

Geschrieben von: / Posted by: Dieter Buerssner at 14 November 2001 19:55:00:
Als Antwort auf: / As an answer to: Re: from WB documantation... geschrieben von: / posted by: Miguel A. Ballicora at 14 November 2001 17:16:15:

A djgpp compile worked perfect for me in my machine and it was 25% faster
than a cygwin compile.
However, it did not work in other machines for reasons
that I do not know. That version was 100% ANSI-C though (so, no ponder of any kind).
IIRC, that select() function is a "dummy" function.

Strange - they produce the same code, but have different libraries, startup code, etc. But library code should not take any time in a chess engine.
In my tests, my prorgams ran at identical speed when compiled with cygwin, Mingw and djgpp.
Do you know, if it worked in text mode? In Standard-C, how did you do the timing? With time()? Would probably be not easy for fast games.
While many Unixish functions are stubs in djgpp, select() is real It uses some DOS/BIOS interrupts. It works even for stdin as a kbhit(), but I did not test with pipes.
Regards,
Dieter

[Miguel A. Ballicora]

Geschrieben von: / Posted by: Miguel A. Ballicora at 14 November 2001 22:11:09:
Als Antwort auf: / As an answer to: Re: from WB documantation... geschrieben von: / posted by: Dieter Buerssner at 14 November 2001 19:55:00:

A djgpp compile worked perfect for me in my machine and it was 25% faster
than a cygwin compile.
However, it did not work in other machines for reasons
that I do not know. That version was 100% ANSI-C though (so, no ponder of any kind).
IIRC, that select() function is a "dummy" function.

Strange - they produce the same code, but have different libraries, startup code, etc. But library code should not take any time in a chess engine.
In my tests, my prorgams ran at identical speed when compiled with cygwin, Mingw and djgpp.
Do you know, if it worked in text mode?
While many Unixish functions are stubs in djgpp, select() is real It uses some DOS/BIOS interrupts. It works even for stdin as a kbhit(), but I did not test with pipes.
Regards,
Dieter

Yes. Some people reported that it did not work with winboard but it did
as a console.
In Standard-C, how did you do the timing? With time()? Would probably be not easy for fast games.
I did it with clock(). I know it is not correct, but it had a much better
resolution (close to centiseconds, I think) than time(). IIRC, time() had
a resolution of a whole second. As long as I did not have any other program
running at the same time, clock() did not give me headaches. Maybe it is
not implemented in djgpp as it should?
I came to the conclusion that there is no efficient way to do this portably.
I am using now what windows provide.
As many times happen, I was wrong. I might have been confused with some other
unix function.