A result about the power of geometric oracle machines (Q1177163)

The author presents a new simpler proof of a result which was contained already in his doctoral thesis (1986): Geometric constructions by compass and ruler in the Euclidean plane (with a given Cartesian coordinate system) which are free of loops (i.e. for which a constant upper bound of the number of steps is known) and in which only such tests occur whether an arbitrary point is the point \((0,0)\) (equivalent is: whether two points are equal) are characterized by algebraic description of the functions \(F\) which can be realized by such algorithms: \(F\) is constructible iff its domain \(S\) can be decomposed into finitely many subsets \(S_ 1,\dots,S_ k\) such that each of these sets is the intersection of finitely many algebraic surfaces and finitely many complements of such and the restriction \(F/S_ i\) of \(F\) on such a subset \(S_ i\) may be described by rational functions with integer coefficients, giving the coordinates of the constructed points from the coordinates of the given points. The surprising elimination of occurring square roots is mainly caused by the condition that the results of the construction are not permitted to depend on the steps of choice of one of the two points of intersection of two circles or a circle and a straight line. From the point of view of classical theory of geometric constructions the result is not at all trivial and typical methods of the general theory of computations are used to prove it. However I think it belongs to the theory of geometric constructions more than to any sort of computer sciences. There is a little error in the example on page 235: One should take the address \(P_ 7\) instead of \(P_ 6\).