[pgn]
[Event "DOCCC 1998"]
[Site "Alphen aan den Rijn"]
[Date "1998.11.21"]
[Round "1"]
[White "Rookie 2.0+"]
[Black "CilkChess"]
[Result "0-1"]
[ECO "A57 Benko gambit half accepted"]
1. c4 Nf6 2. d4 c5 3. d5 b5 4. cxb5 a6 5. e3 g6 6. Nc3 d6 7. Nf3 Bg7 8. Be2 O-O 9. a4 axb5 10. Bxb5 Ba6 11. Bxa6 Nxa6 12. O-O Nb4 13. e4 c4 14. Kh1 Nd3 15. Qe2 Qb6 16. Ne1 Nxc1 17. Rxc1 Qb4 18. Ra1 Rfb8 19. Ra2 Qb3 20. f3 Nd7 21. Rf2 Bd4 22. Rf1 Nc5 23. a5 Bxc3 24. bxc3 Na4 25. Qd2 Nxc3 26. Ra1 Na4 27. Nc2 c3 28. Qf2 Qc4 29. Qd4 Rc8 30. Rfb1 Qxd4 31. Nxd4 Rc4 32. Nb5 Nc5 33. Ra2 c2 34. Rc1 Nd3 35. Rcxc2 Rxc2 36. Rxc2 Rxa5 37. h4 Rxb5 38. Kh2 f5 39. Rc8+ Kg7 40. Re8 Rb7 41. Kg3 fxe4 42. fxe4 Kf6 43. Rh8 h5 44. Rg8 Nc5 45. e5+ dxe5 46. Rf8+ Kg7 47. Rd8 Ne4+ 48. Kf3 Nd6 0-1
[/pgn]
This is a picture of us after the game. From left to right: Vincent Diepeveen, Don Dailey, Aske Plaat, an unknown spectator (not MIT team member), and me.

Don's approach to chess programming was already much more professional than most of us. The idea of testing by playing many automated games was not commonly done by amateurs. Don explained the difficulties he had interfacing with commercial programs. Nothing like UCI existed and all programs came with their own GUI. His solution to play them anyway: frame grab the display from the video memory and detect where the pieces are drawn. Send back your own moves by faking keyboard interrupts. Absolutely brilliant.
Don will be severely missed.


