The paper examines the structure obtained by abstracting from the conventional (manifold) representation of relativistic space-time the concept of an event-set equipped with two partial orderings, whose counterparts are the notions ‘causally precedes’ and ‘chronologically precedes in the history of some observer’.

A prescription is given for attaching to a space-time M , subject only to a causality condition, a collection of additional ‘ideal points’. Some of these represent ‘points at infinity’, others ‘singular points’. In particular, for asymptotically simple space-times, the ideal points can be interpreted as the boundary at conformal infinity. The construction is based entirely on the causal structure of M , and so leads to the introduction of ideal points also in a broad class of causal spaces. It is shown that domains of dependence can be characterized in terms of ideal points, and this makes possible an extension of the domain-of-dependence concept to causal spaces. A suggestion is made for assigning a topology to M together with its ideal points. This specifies some singular-point structure for a wide range of possible space-times.

In this paper we study those regular fenestrations (as defined by Kronheimer in [3]) that are obtained from a tiling of a topological space. Under weak conditions we obtain that the canonical grid is also the minimal grid associated to each tiling and we prove that it is a T0-Alexandroff semirregular trace space. We also present some examples illustrating how the properties of the grid depend on the properties of the tiling and we pose some questions. Finally we study the topological properties of the grid depending on the properties of the space and the tiling.

Do there exist topologies U \mathcal {U} and V \mathcal {V} for the set R R of real numbers such that a function f f from R R to R R is smooth in some specified sense (e.g., differentiable, C n {C^n} , or C ∞ {C^\infty } ) with respect to the usual structure of the real line if and only if f f is continuous from U \mathcal {U} to V \mathcal {V} ? We show that the answer is no.

No abstract