123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960616263646566676869707172737475767778798081828384858687888990919293949596979899100101102103104105106107108109110111112113114115116117118119120121122123124125126127128129130131132133134135136137138139140141142143144145146147148149150151152153154155156157158159160161162163164165166167168169170171172173174175176177178 |
- .TH Chess GNU
- .SH NAME
- Chess \- The Technology Chess Program for GNU Unix
- .SH SYNOPSIS
- .B Chess
- [
- .BI \-n
- ]
- .SH DESCRIPTION
- .I Chess
- plays a game of chess against the user. Or it plays against
- itself. Or it referees a game.
- .PP
- .I Chess
- has two display modes. The first mode is simply a normal
- mode that you could use with a terminal. The second mode
- is a fancy display mode you can use with a SUN workstation.
- To use the former, simply type 'gnuchess'. To use the
- latter, simply type 'chesstool gnuchess' on a SUN
- workstation where 'chesstool' is installed. In the
- latter example, the argument should be the path
- specifying where to find the gnuchess binary.
- .PP
- The following documentation assumes you are in
- the first mode (e.g. normal mode using a regular
- terminal). If not, then you should read the chesstool documentation.
- .PP
- For help once in
- .I Chess
- type a question-mark. To type in your move, use the
- notation "e2e4" where the first letter-number pair
- indicates the origination square and the second
- letter-number pair indicates the destination square.
- The letter indicates the column with the left-most
- column being "a" and the right-most column being "h".
- The number indicates the row, the first row (White's
- first rank) being "1" and the last row (Black's
- first rank) being "8". To castle, type the origin
- square of the king and the destination square of the
- king, just as you would do for a regular move.
- .PP
- The "bd" command prints what the current board position
- looks like. You may type this to see what the board
- looks like after the computer moves.
- .PP
- The "book" command compiles the opening book into
- dbm(3) format. This book is then consulted
- whenever the computer makes a move. If the current
- position is in the book, then the suggested move
- associated with that position will be made as
- the computer's move. Note that this command is a
- maintenance command, usually used only once per
- site per book.
- .PP
- The "depth" command allows the user to change how many
- moves ahead the computer looks. Normally, it looks ahead
- three half-moves, examining every move for each side.
- Then, it examines all captures for both sides for another
- two half-moves. "Depth" changes the first of these values.
- There is an upper boundary of five half-moves as set by
- the quiescence search which searches captures a little
- deeper to get better tactical play.
- .PP
- The "enter" command causes the current game played thus
- far to be entered in the human-readable version of the
- opening book. If no current game is in progress, this
- command requests the name of a file from which to read
- games. These are then added to the binary database.
- Since the versions of the two files are slightly different,
- in the former case, you must additionally type "book"
- to get the new game fully entered in the book.
- Also, the "enter" command tries to mail your new additions
- to the book maintainers
- so that your contributions can be added to the master copy
- of the book. If you are not on our network, you should send
- us your new additions via some other method.
- .PP
- "Fill" allows the user to specify a completely different
- board position. Input is based on Forsythe notation.
- For example, the opening position in Forsythe notation
- is "rnbqkbnrpppppppp8888PPPPPPPPRNBQKBNR+". Capital
- letters indicate a White piece, lower-case Black. A plus means
- it is White's move, a minus means Black's.
- .PP
- "History" and "historyf" list the game as played so far
- to the terminal and a file respectively. The file is
- assumed to be "GAMES/chXXXXXX" where XXXXXX is a random
- number and GAMES is a subdirectory. If GAMES does not
- exist, an error message is reported. The equivalent of
- "historyf" is done after every move automatically to
- record the game thus far.
- .PP
- "Legals" shows legal moves for the current position along with
- the rating for each move based on a positional presort.
- .PP
- "Neither" instructs the program to play neither side, that is,
- to simply act as a referee.
- .PP
- "Parallel" enables or disables parallelism, assuming it has
- been available when your version of the program was configured.
- In its enabling mode, this command lists the names of the
- processors which will be used along with their machine architecture.
- .PP
- "Reset" resets the board to the starting position.
- .PP
- "Quit" exits the game.
- .PP
- "Read" restores a game as if you were still playing it.
- The game must be in the format as written by "historyf" or
- as automatically recorded by the program itself after every
- move.
- .PP
- "Self" causes the program to play against itself.
- .PP
- "Static" causes a static evaluation to be done for the
- current position. A static evaluation is based on material
- difference only. Positional considerations are handled
- by the ply-1 positional presort.
- .PP
- "Switch" causes the program to move, whether or not it
- is the program's turn to do so. Continually typing
- "switch" is equivalent to typing "self".
- .PP
- "Test-moves" initiates a test of the speed of the move
- generator. An initial series of move generations is done for the
- opening position and the timing speed is reported.
- Then, for each of ten stored test positions, a series of
- move generations is done and the timing speed is reported
- for each one. Finally, the program averages across those
- ten runs and reports the overall average.
- .PP
- "Test-search" uses the ten stored test positions to
- actually conduct a search at the current depth to
- choose a move for each position.
- .PP
- "Undo" undoes the last move whether it was the computer's
- or the human's. You may also type "remove". This is equivalent
- to two "undo's" (e.g. retract one move for each side).
- .PP
- The flag-option on the command line allows specification
- of how deep to search in half-moves as with 'Chess -3'
- to search three half-moves ahead. Capture searches are
- normally carried out a few half-moves further than the
- regular full search specified with this option. This
- is known as the quiescence search and it usually is
- conducted to 6 ply. This option is identical to the "depth"
- command once in the program.
- .SH BUGS
- .PP
- En passant is not currently implemented. Also, promotion
- to pieces other than queens is disallowed. Checks and
- checkmates are not detected in the tree-search and
- are not handled as "forcing" variations.
- .PP
- There are other bugs. Suggestions for improvements
- and caveats are contained in the files README and TODO
- which come with this distribution.
- .SH AUTHOR
- .nf
- Stuart Cracraft Stuart Cracraft
- P.O. Box 13123 UCLA, Dept. of Mathematics
- Torrance, Ca. Los Angeles, Ca.
- 90503 90024
- (213) 214-1136 (213) 825-9040
- .fi
- .SH AUTHOR'S COMMENT
- .PP
- This software is being made available by the Free Software
- Foundation under the restrictions described in its license
- agreement which accompanies this distribution. This software,
- its sources, binaries, documentation and all associated parts
- are copyright (C) 1986 by the Free Software Foundation, Inc.
- .SH SEE ALSO
- .nf
- chesstool(6)
- dbm(3)
- .fi
|