[computer-go] Assistance needed

[Nick Wedd]

In message 
<6a9f80570609030655p31b4a766t9b1e560c2cfbba98 at mail.gmail.com>, Rahul S K 
<rahul.sk.iitb at gmail.com> writes

>             I am a junior undergraduate from the Indian Institute of
>Technoloy(IIT), Bombay, India and am a part of the Techfest organising
>team. Techfest is IIT Bombay's Annual International Science and
>Technology festival.(www.techfest.org)
>           This year, at Techfest, one of the section of competitions
>called "wired logic" deals with computer science related competitions.
>We are exploring the possibility of holding an online code v/s code GO
>competition for students and am currently stuck up on some of the
>following points, hope someone can assist me :
> 1. Since this is going to be a competition mainly for junior
>undergrads who are going to put in about 3 months of efforts, learning
>the entire protocol for computer go seems , infeasible.

I think the protocol is fairly simple, particularly compared with what 
you propose next.

> However a
>simple C++ / Java program that takes the board scenario as an input
>(possibly a 19X19 array) , and gives out the next move is what I have
>in mind.

>2. However in such a code vs code competition, the need to have a
>"supervising" program that can understand the territories, life-death
>situations and end game scoring complexitites is a big hurdle that I
>hope to cross... any help in this regard would be most welcome...any
>program already available that you possibly know of would be most
>appreciated.

I don't think that such a supervising program is within anyone's 
ability.

If you want to make this competition work, I suggest getting the 
competing programs to continue play until all "dead" stones have been 
removed from the board.  And, use Chinese (area) scoring rather than 
Japanese (territory) scoring.  Once there are no dead stones left on the 
board, a supervising program should be practicable.

Nick
-- 
Nick Wedd    nick at maproom.co.uk