howard@sci.UUCP (Howard Landman) (01/21/85)
The bibliography posted recently missed at least one reference, namely: "Programs for the Japanese Game of Go", chapter 12 in Uhr, Leonard, _P_a_t_t_e_r_n _R_e_c_o_g_n_i_t_i_o_n, _L_e_a_r_n_i_n_g, _a_n_d _T_h_o_u_g_h_t, Prentice-Hall 1973 This is basically an analysis of Zobrist's program from the viewpoint of pattern recognition, but it includes skeleton code for the program (in an obscure variant of SNOBOL called EASEy). I have the book and can answer questions about it. I think there might be a few others missing. I'll check my bibliography at home and report if there are. How about a project to get all of the existing Go literature in computer- readable form and post it on the net (getting the appropriate permissions first, of course). I'd be willing to do the Benson article on life and death since there are several mistakes in his formalization of the rules and I've got detailed notes on them. For example, I believe that on a 3 by 6 Go board, his rules would make it illegal for White to play the only winning move in this position: B B B B W + + B W + B W B B B B W + This is due to an erroneous formulation of the "No suicide" rule. Most of his conclusions are valid, and the proofs can be fixed up with minor effort, but the naive reader might try using his postulates verbatim in a Go program, with the result that the program wouldn't be playing Go at all! Howard A. Landman sci!howard