Abstract. This survey paper describes the role of splines in geometry and topology, emphasizing both similarities and differences from the classical treatment of splines. The exposition is non-technical and contains many examples, with references to more thorough treatments of the subject.The goal of this survey paper is to describe how splines arise in geometry and topology. Geometric splines usually appear under the name GKM theory after Goresky-Kottwitz-MacPherson, who developed them to compute what is called the cohomology ring of a geometric object. Geometers and analysts ask many of the same questions about splines: what is their dimension? can we identify a basis? can we find explicit formulas for the elements of the basis? However geometric constraints can change the tone of these questions: the splines may satisfy various symmetries or have a basis satisfying certain conditions. And some questions are specific to geometric splines: geometers particularly care about the multiplication table with respect to a given basis.In Section 1 we discuss GKM theory, followed in Section 2 by some important families of geometric examples, some of which are well-known to students of analytic splines and some of which may be useful in future. Section 3 sketches some techniques that are natural from the perspective of a geometer/topologist, including Morse flows and symmetries that come from geometric representation theory. Finally in Section 4 we generalize splines to a more abstract ring setting, both as a useful conceptual framework and because it provides new combinatorial tools. This paper is targeted at researchers in geometric design, especially those with an analytic background. Our aim is to give an overview of theoretical tools and techniques from geometry and topology; we often illustrate concepts by example and refer to the literature for details on technical aspects. A reader with a different mathematical perspective may be interested in surveys like [27,35,36].
GKM theoryCohomology is an algebraic gadget associated to a geometric object X that encodes various properties of X. Among other things, the cohomology of X indicates the dimension of X, the number of connected components (how many separate pieces X has), how many holes X has, whether X has singularities, and how different subspaces of X intersect.We could treat very general kinds of geometric objects but for simplicity in this survey we take X to be a compact complex manifold. When we say that cohomology is an "algebraic gadget" the most general interpretation is that cohomology is a ring. In the cases of interest here, the cohomology ring is actually an algebra, namely a vector space in which one can multiply vectors. The technical condition we assume is that cohomology has coefficients in Q, R, or especially C.In fact we will consider an enhanced version of cohomology called T -equivariant cohomology. Equivariant cohomology has strictly more information than ordinary cohomology yet surprisingly can be easier to compute. In our case T is a torus, ...