Re: Intelligence

[Compgo123@xxxxxxx]

 I think some of the discussions are getting very close as how to write a 
good Go program.  

I see no reason why a computer program (not necessarily run on a pc) cannot 
exceed 6 dan. Serge Boisse has some good points. But a computer's capacity in 
memeory far exceeds that of a human. I see no inherent obstacle to 
incorporate all available Go knowledge by a computer. I think the problem 
lies in the creativity and interactions bewteen knowledges. A top pro player 
cannot win a game if he only play conventional moves. He wins often because 
that he played an unexpected, unconventional new move. However, this weakness 
can be made up by using the search capability of a computer.

I have being wondering too that if a top pro player has deeper knowledge 
about the GO game than what we usually read from a Go book. It's possible, 
because orientals has the tradition and tendency to do so. However, there 
seems lack of indication about this.

Talking about Go knowledge, some people think that most of the books we are 
reading (from Ishi press)  represent the Japanese Go theories in the 70's. Of 
course, the bulk of it remains valid as to today. 


Dan Liu