SOME PROBLEMS OF AUTOMATA CONSTRUCTION (A VIEWPOINT)

The functioning of computers is based on mathematical logic. Hence any substantial progress in computer construction must come from a brilliant idea in logic. This probably entails a large increase of the number of basic units and of their interconnections. This is, however, feasible in practice provided miniaturization could be much more developed than in existing machines, but this is not possible unless a new type of physics could be established, viz. the physics of highly inhomogeneous or asymmetric systems. A more advanced miniaturization probably entails a decrease of the reliability of the basic units. The use of large numbers of basic units of a relatively low reliability requires new systems of error and failure correction, which must circumvent the use of prohibitively large numbers of basic units, particularly since finding a faulty unit in a million or so in them is like looking for a needle in a haystack. The solution of these problems does not seem to be in sight.