Tony M wrote:

I made an earlier post recommending that Sanny look at the source of
the Java version of TCSP for an idea on implementing a time control so
that a program plays in exactly x number of seconds. I made the
recommendation because TSCP handles it in a simple way, and in only
about a dozen lines of code.

In Java/chess programming speak (help bot, you may want to skip this
paragraph, it's hellaciously geeky), TSCP puts the entire root search

(root search: this is what rabbits do when looking for carrots)

in a try block

(try block: a block with precisely three sides)

, checks for a timeout

(timeout: what you put your kids in when they misbehave)

every 1024 nodes

(nodes: similar to nematodes, but without the emats)

(both in the alpha/beta

(Alpha beta is a large grocery store chain.)

and quiescence search), throws an exception

(An "exception" is any pitch which is not a fastball, curve, or
screwball)

when a timeout
is reached, and plays the best move found so far when it catches the
exception

(easier said than done!)

, after taking back whatever moves that it made

(see for example, Kasparov vs. J. Polgar)

during the search when it timed out. It might not be the most
elegant way to do things, but it works well enough.

(Point taken. Kasparov 1, Polgar 0)


Just to make it clear, it's not my intent to belittle Sanny and his
efforts. I know from personal experience how hard it is to write a
bug-free chess program, and how bugs can sap the strength of a
program. He's improved the quality of play over the last little
while, to the point where it plays reasonably well for a Java applet.
I'm a fan of all chess programs, big and small, and would just like to
see Sanny's program be exactly what he advertises it to be.

Things are tough all over. Sanny needs people to buy
games in order to finance further improvements, but they
are unlikely to do so unless and until these improvements
are implemented. Zugzwang. (That's geek-speak for "any
move loses").

-- help bot