New open source engine called "Pawny"

[Graham Banks]

Hello, Mr. Banks. My name is Mincho Georgiev and my engine is called "Pawny".

The website is located on:
http://pawny.netii.net/

P.S. You are the first to know for my release. I would be grateful if you share that information with others as well!
Many thanks in advance!

[Christopher Conkie]

Hello, Mr. Banks. My name is Mincho Georgiev and my engine is called "Pawny".

The website is located on:
http://pawny.netii.net/

P.S. You are the first to know for my release. I would be grateful if you share that information with others as well!
Many thanks in advance!

http://www.000webhost.com/admin-review

That was quick.....



Christopher

[Spock]

I got that too, but the page is now back

[bnemias]

Compiled on GNU/Linux, but only after substituting "sort.o" with "see.o" in the Makefile.

[gerold]

Hello, Mr. Banks. My name is Mincho Georgiev and my engine is called "Pawny".

The website is located on:
http://pawny.netii.net/

P.S. You are the first to know for my release. I would be grateful if you share that information with others as well!
Many thanks in advance!

Thanks for the post Graham. Thanks Mincho for the engine.
It works in Arena.

Best,

Gerold.

[Dann Corbit]

Hello, Mr. Banks. My name is Mincho Georgiev and my engine is called "Pawny".

The website is located on:
http://pawny.netii.net/

P.S. You are the first to know for my release. I would be grateful if you share that information with others as well!
Many thanks in advance!

It's radically different now, but I think pawny began its life as tscp. For example:

C:\pgn\winboard-engines\pawny\src>grepcarl pawn_mat \dannfast\e_drive\tscp\*.c
\dannfast\e_drive\tscp\eval.c ( 118): int pawn_mat[2]; /* the value of a side's pawns */
\dannfast\e_drive\tscp\eval.c ( 130): /* this is the first pass: set up pawn_rank, piece_mat, and pawn_mat. */
\dannfast\e_drive\tscp\eval.c ( 137): pawn_mat[LIGHT] = 0;
\dannfast\e_drive\tscp\eval.c ( 138): pawn_mat[DARK] = 0;
\dannfast\e_drive\tscp\eval.c ( 144): pawn_mat[color] += piece_value[PAWN];
\dannfast\e_drive\tscp\eval.c ( 159): score[LIGHT] = piece_mat[LIGHT] + pawn_mat[LIGHT];
\dannfast\e_drive\tscp\eval.c ( 160): score[DARK] = piece_mat[DARK] + pawn_mat[DARK];

C:\pgn\winboard-engines\pawny\src>grepcarl pawn_mat *.c
eval.c ( 58): int pawn_material[2];
eval.c ( 94): pawn_material[0] = pawn_material[1] = 0;
eval.c ( 122): pawn_material[color] += P_VALUE;
eval.c ( 264): score[WHITE] += material[WHITE] + pawn_material[WHITE];
eval.c ( 265): score[BLACK] += material[BLACK] + pawn_material[BLACK];
eval.c ( 453): ||pawn_material[color] > pawn_material[color ^ 1])

The author should at least acknowlege that heritage.

IMO-YMMV

[Graham Banks]

Hello, Mr. Banks. My name is Mincho Georgiev and my engine is called "Pawny".

The website is located on:
http://pawny.netii.net/

P.S. You are the first to know for my release. I would be grateful if you share that information with others as well!
Many thanks in advance!

It's radically different now, but I think pawny began its life as tscp. For example:

C:\pgn\winboard-engines\pawny\src>grepcarl pawn_mat \dannfast\e_drive\tscp\*.c
\dannfast\e_drive\tscp\eval.c ( 118): int pawn_mat[2]; /* the value of a side's pawns */
\dannfast\e_drive\tscp\eval.c ( 130): /* this is the first pass: set up pawn_rank, piece_mat, and pawn_mat. */
\dannfast\e_drive\tscp\eval.c ( 137): pawn_mat[LIGHT] = 0;
\dannfast\e_drive\tscp\eval.c ( 138): pawn_mat[DARK] = 0;
\dannfast\e_drive\tscp\eval.c ( 144): pawn_mat[color] += piece_value[PAWN];
\dannfast\e_drive\tscp\eval.c ( 159): score[LIGHT] = piece_mat[LIGHT] + pawn_mat[LIGHT];
\dannfast\e_drive\tscp\eval.c ( 160): score[DARK] = piece_mat[DARK] + pawn_mat[DARK];

C:\pgn\winboard-engines\pawny\src>grepcarl pawn_mat *.c
eval.c ( 58): int pawn_material[2];
eval.c ( 94): pawn_material[0] = pawn_material[1] = 0;
eval.c ( 122): pawn_material[color] += P_VALUE;
eval.c ( 264): score[WHITE] += material[WHITE] + pawn_material[WHITE];
eval.c ( 265): score[BLACK] += material[BLACK] + pawn_material[BLACK];
eval.c ( 453): ||pawn_material[color] > pawn_material[color ^ 1])

The author should at least acknowlege that heritage.

IMO-YMMV

From the readme file:

Many Thanks to:
----------------
Tom Kerrigan, who send me the old sources of tscp, and help me that way
to understand the array of goals between the different versions!
The helpfull chess comunity from talkchess and winboard forums!

[Spock]

It's radically different now, but I think pawny began its life as tscp. For example:

The author should at least acknowlege that heritage.

IMO-YMMV

He sort of does, in the read-me file

[Dann Corbit]

Hello, Mr. Banks. My name is Mincho Georgiev and my engine is called "Pawny".

The website is located on:
http://pawny.netii.net/

P.S. You are the first to know for my release. I would be grateful if you share that information with others as well!
Many thanks in advance!

It's radically different now, but I think pawny began its life as tscp. For example:

C:\pgn\winboard-engines\pawny\src>grepcarl pawn_mat \dannfast\e_drive\tscp\*.c
\dannfast\e_drive\tscp\eval.c ( 118): int pawn_mat[2]; /* the value of a side's pawns */
\dannfast\e_drive\tscp\eval.c ( 130): /* this is the first pass: set up pawn_rank, piece_mat, and pawn_mat. */
\dannfast\e_drive\tscp\eval.c ( 137): pawn_mat[LIGHT] = 0;
\dannfast\e_drive\tscp\eval.c ( 138): pawn_mat[DARK] = 0;
\dannfast\e_drive\tscp\eval.c ( 144): pawn_mat[color] += piece_value[PAWN];
\dannfast\e_drive\tscp\eval.c ( 159): score[LIGHT] = piece_mat[LIGHT] + pawn_mat[LIGHT];
\dannfast\e_drive\tscp\eval.c ( 160): score[DARK] = piece_mat[DARK] + pawn_mat[DARK];

C:\pgn\winboard-engines\pawny\src>grepcarl pawn_mat *.c
eval.c ( 58): int pawn_material[2];
eval.c ( 94): pawn_material[0] = pawn_material[1] = 0;
eval.c ( 122): pawn_material[color] += P_VALUE;
eval.c ( 264): score[WHITE] += material[WHITE] + pawn_material[WHITE];
eval.c ( 265): score[BLACK] += material[BLACK] + pawn_material[BLACK];
eval.c ( 453): ||pawn_material[color] > pawn_material[color ^ 1])

The author should at least acknowlege that heritage.

IMO-YMMV

From the readme file:

Many Thanks to:
----------------
Tom Kerrigan, who send me the old sources of tscp, and help me that way
to understand the array of goals between the different versions!
The helpfull chess comunity from talkchess and winboard forums!

I had downloaded the source files and not the binaries.

[Mincho Georgiev]

After short considering i decide to post my answer, sent by email to Dan Corbit. Regarding the above example, i don't think that is much indicative, but anyway.

It is clear that TSCP was at least studied in the generation of your chess engine

On Fri, Sep 4, 2009 at 7:46 AM, Dann Corbit <DCorbit@connx.com> wrote:

I think your release notes should credit inspiration from TSCP.

Mr. Corbit,

Initially I didn’t knew how to respond due to the fact that I didn’t realize immediately
if your remark is expressing disapproval, or the case is the opposite,
because personally for me, the comparison between my code and TSCP’s is a compliment.
However, I’ll treat it as disapproval and I’ll continue.
In case you didn’t read the entire “readme.txt”, please, do that.
Actually, Tom is the only one, personally mentioned at “credits”.
That doesn’t mean though that Pawny was “inspired” from TSCP, nor from any
other particular program.
In case you did read the “readme.txt” and made your remark after all,
I feel responsible to answer and explain a couple of things.
Some time ago, Tom Kerrigan kindly respond and send me the code for the old versions of TSCP.
As reason for my request to him back then I’ve pointed my personal need
of more simplified version than 1.81 since I wanted to try different board representations,
combined with his code and all that for educational purposes. However, after certain period of reading his code,
I decided not to use it, but to start writing the engine, that
I was already had plans for without experimenting with his code.
Regardless of my decision, Tom’s code help me a lot mainly to understand some basic sequence of objectives to follow.
I don’t deny at all, that there are certain ideas (not code) from TSCP that I am using.
For example, the pawn structure representatives for evaluation, but probably 60%
of the programs without pawn hashing are using the same method as well! There are
far more educated and better programmers that me, who can point the originator
of that method.
I think that almost anything else comes down to programming (coding) style and I don’t
see anything wrong (I’m even flattered) , if you see something in common between
mine and Tom’s style, in spite of mine is not even as half as good.
In these thoughts, I would like to continue and to point some other possible similarities
with other programs. And such exists, I believe. That doesn’t mean that Pawny was
inspired by any of these programs either.
For example, the constants I’m using for piece values. Glaurung and probably Diablo are
using the exact same names (P_VALUE, K_VALUE, e.t.c.). Apparently, the data type
names, ending with “_t” as well. Diablo checks for input on every 0x1FFF nodes.
Pawny - same node count. I was using ideas from Bruce Moreland’s site, as I’m sure,
many other programmers.
Is it possible however, to have some personal ideas in my code, useful for others
and I will be glad if someone use them.
Pawny was entirely written from scratch and there was no copy->paste from any other
program’s code except the bioskey procedure from Olithink and Beowulf.
In conclusion, I would like to say two things:
First – I sincerely regret that I had to write this letter,
but maybe this is nothing uncommon among the chess programming community.
I’m just not that experienced to know.
Second – maybe there will be some people, for whom Pawny is just the next “bloataware of code”,
or asking “Why another”, especially if it’s not strong or competitive enough.
I just could say the following:
Is there less chess players around the world than chess programs?
Or, would it be a bad idea for a small kid to start studying chess just because
there are already too many chess players ?
I don’t think so !


Best Regards!
Mincho Georgiev.