[gnugo-devel] GNU Go Development Idea

[Bowdridge, Steven James]

Hello GNU Go Developers,

I am an undergraduate student studying Computer Science at the University of New
Brunswick, Canada.  I am graduating in the spring, and have been working on a 
thesis
paper relating to GNU Go.  

My paper discusses approaches taken to programming Go engines, in comparison to 
other
games, etc.  For the programming component of my project, I considered 
different ideas,
and settled on adding pondering (thinking on opponent's time) to the engine.

I see this idea has been tossed around on the gnugo-dev mailing lists 
previously, and
would like to know if there are any pros and cons, or analysis done on the idea 
of
pondering.  From my research I understand some commercial Go engines have 
implemented
pondering, so it may have some benefits.

Currently, the modification I have is very rough, but I would be happy to give 
the
source code to anyone who wants it.  It basically ponders the top 10 expected
responses, and checks back for user input between pondering each move.  If the 
opponent
has made one of the moves that has been pondered, GNU Go doesn't need to 
consider a
response, and instantly responds.  The idea is to save some time, since from
what I understand GNU Go lowers its skill level when running low on time. No 
threads or
extra processes are spawned here.

Any input you have on this idea would boost my thesis enormously. Whether 
positive or
negative, I would love to know the developers' thoughts on the idea.  I would 
be happy
to correspond my findings with anyone interested in the topic.  

Thank you,

Steven
University of New Brunswick