[Alan Turing]
p.14 If one can explain quite unambiguously in English, with the aid of mathematical symbols if required,
how a calculation is to be done, then it is always possible to programme any digital computer to do that calculation, provided
the storage capacity is adequate.
p.15 One does not follow all the continuations of play, but one follows some of them. One
does not follow them until the end of the game, but one follows them a move or two, perhaps more. Eventually a position seems,
rightly or wrongly, too bad to be worth further consideration, or (less frequently) too good to hesitate longer over.
p.16 Note that no "analysis" is involved in position-play evaluation. This is to reduce the amount
of work done on deciding the move.
p.17 If I were to sum up the weakness of the above system in a few words I would describe it as a caricature
of my own play. It was in fact based on an introspective analysis of my thought processes when playing, with considerable
simplifications.
p.17 one can see that it would be quite possible to programme the machine to try out variations in its method
of play (e.g. variations in piece value) and adopt the one giving the most satisfactory results.
[Alan Newell]
p.18 The modern general-purpose computer can be characterized as the embodiment of a three-point philosophy:
(1) There shall exist a way of computing anything computable; (2) The computer shall be so fast that it does not matter how
complicated the way is; and (3) Man shall be so intelligent that he will be able to discern the way and instruct the
computer.
[A.D. de Groot]
p.171 Calculating a variant and working out a plan are, in fact, operations which in any experienced
chess player are completely controlled by routine. These operations, envisaging a position included, are executed so often
in practice that the process runs almost automatically
p.171 Which factors determine the order in which possibilities or variants are to be investigated?
[JLJ - a great question.]
p.172 In general, priority rules are sensible and understandable. Thus, in most situations the investigation
of calmer moves which maintain the tensions and the status quo is not economical until it is known that there is
no immediate advantage to be gained of threat to be parried.
p.173 As always, when we talk about "seeing" in chess, it is the perception of the relations between
the visible objects and not the perception of the objects (board and pieces) themselves. [JLJ - another great idea]
[Herbert Simon and William Chase]
p.175 The MATER theory is an application to the chess environment of a more general theory of problem solving
that employs heuristic search as its core element (Newell el at, 1972). The MATER theory postulates that problem solving in
the chess environment, as in other well-structured task environments, involves a highly selective heuristic search through
a vast maze of possibilities. Normally, when a chess player is trying to select his next move,
he is faced with an exponential explosion of alternatives... The MATER theory postulates that humans don't consider
moves at random. Rather, they use information from a position and apply some general rules (heuristics) to select a small
subset of the legal moves for further consideration.
p.176 a theory of problem solving in chess that does not include perceptual processes cannot be
an adequate theory - cannot explain the superior ability of the strong player to choose the right moves.
p.177 the Russian experiments confirm the existence of an initial "perceptual phase," earlier hypothesized
by de Groot, during which the players first learn the structural patterns of the pieces before they begin to look for a good
move in the "search phase" of the problem-solving process.
[G.W. Baylor and H.A. Simon]
p.211 This ends what might be called the "static" perceptual relations on the chess-board. What follows
is a bundle of basic routines what attempts to provide "dynamic" perceptual relations to the program.
