We improve the intuition and some ideas of G. D. Faulkner and M. C. Vipera presented in the article "Remainders of compactifications and their relation to a quotient lattice of the topology" [Proc. Amer. Math. Soc. 122 (1994)] in connection with the question about internal conditions for locally compact spaces X and Y under which /3X \ X = f)Y \ Y or, more generally, under which the remainders of compactifications of X belong to the collection of the remainders of compactifications of Y. We point out the reason why the quotient lattice of the topology considered by Faulkner and Vipera cannot lead to a satisfactory answer to the above question. We replace their lattice by the qoutient lattice of a new equivalence relation on a Wallman base in order to describe a method of constructing a Wallman-type compactification which allows us to deduce more complete solutions to the problems investigated by Faulkner and Vipera.
INTRODUCTIONAll topological spaces considered below will be locally compact and HausdorfF. The word "space" will be used to refer to such, a topological space, unless otherwise specified.For a space X, the symbol K.(X) will denote the lattice of all compactifications of X, while TZ(X) will stand for the collection of all remainders of X, that is, TZ(X) -{aX \ X : aX £ )C(X)} if we do not distinguish between homeomorphic spaces.Although it is generally known that non-homeomorphic spaces can have homeomorphic remainders of their Cech-Stone compactifications, the nature of such spaces is not well understood. For instance, how to compare such a trivial space as the space N of positive integers with a pseudocompact non-normal space A = /3E \ (/?N \ N) (see [10, 6P])? But the reason for which /3N\ N = /3A \ A must surely be hidden somewhere in the internal structures of the topologies of N and A.The question of when a space K can be homeomorphic to /3X \ X is also nontrivial. Let us mention that, for instance, the statement "every Parovicenko space of weight 2 W is homeomorphic to /3N\N" is equivalent to the continuum hypothesis (see [8]). Furthermore, by Magill's theorem (see [6, 7.2]), the inclusion TZ(X) C ~R{Y)