Reconstructing Turing's "paper machine"

EasyChair Preprint no. 3

Abstract

It is an amazing fact that the very first chess program in history was written a few years before computers had been invented. It was designed by a visionary man who knew that programmable computers were coming and that, once they were built, they would be able to play chess. The man, of course, was Alan Turing, one of the greatest mathematicians who ever lived. Soon after the war he wrote the instructions that would enable a machine to play chess. Since there was as yet no machine that could execute the instructions he did so himself, acting as a human CPU and requiring more than half an hour per move. A single game is recorded, one in which Turing's "paper machine" lost to a colleague.

We sketch the historical context of Turing’s involvement in chess and then go on to describe how the chess computer experts reconstructed the paper machine to run on a modern day computer. In the process they encountered a problem: the chess engine refused to duplicate all of Turing’s moves as recorded in the historical game. The debugging process, in which computer chess pioneer Ken Thompson was involved, left the programmers baffled. Then someone called Donald Michie, a colleague from Bletchley, who advocated debugging not the program but Turing! “Alan did not care about details; he was interested in general principle.” In this paper we also discuss the points of deviation from the recorded game.

Keyphrases: Alan Turing, Chessbase, computer chess

@Booklet{EasyChair:3,
  author = {Garry Kasparov and Frederic Friedel},
  title = {Reconstructing Turing's "paper machine"},
  howpublished = {EasyChair Preprint no. 3},
  doi = {10.29007/g4bq},
  year = {EasyChair, 2017}}