the study of properties of space preserved under continuous deformation — stretching, bending, twisting — but not tearing or gluing. two shapes are topologically identical if one can be continuously deformed into the other; a coffee cup and a torus are the same, a sphere and a torus are not
the foundational shift: topology replaces distance with openness. a topology on a set $X$ is a collection $\tau$ of subsets — the open sets — satisfying: $\emptyset$ and $X$ are open; arbitrary unions of open sets are open; finite intersections of open sets are open. from these three axioms, without any notion of measurement, the entire structure of continuity, convergence, and connectedness follows
continuous maps and equivalence
a map $f: X \to Y$ is continuous when preimages of open sets are open — this is the intrinsic definition, free of $\varepsilon$-$\delta$ language. a homeomorphism is a bicontinuous bijection: the strongest notion of topological equivalence. spaces that are homeomorphic are indistinguishable by any topological property
weaker equivalences matter too. a homotopy equivalence allows deformation retracts: the circle and the punctured plane are homotopy equivalent but not homeomorphic. homotopy is the study of continuous deformations between maps, organized into homotopy groups $\pi_n(X)$. the fundamental group $\pi_1(X)$ — loops up to continuous deformation — classifies which holes a space has at dimension one
invariants: counting what topology cannot change
topological invariants are quantities preserved by homeomorphism. they are the tools for distinguishing spaces without finding an explicit homeomorphism or proving none exists
the Euler characteristic $\chi = V - E + F$ for a polyhedron, generalized by the alternating sum of Betti numbers: $\chi = \sum_n (-1)^n b_n$. for the sphere $\chi = 2$, the torus $\chi = 0$, the double torus $\chi = -2$
homology groups $H_n(X)$ count $n$-dimensional holes algebraically: $H_0$ measures connected components, $H_1$ measures loops that bound no disc, $H_2$ measures enclosed voids. homology is computable and turns topological questions into linear algebra
cohomology $H^n(X)$ is the dual theory, assigning functions to holes rather than counting chains. cohomology carries a ring structure — the cup product — and connects to differential forms via de Rham's theorem. it is the natural home of sheaf cohomology
sheaves: local-to-global consistency
a sheaf $\mathcal{F}$ on a topological space $X$ assigns data $\mathcal{F}(U)$ to each open set $U$, with consistent restriction maps whenever $V \subseteq U$, satisfying two axioms: locality (sections that agree locally are equal) and gluing (locally consistent sections assemble into a unique global section)
sheaves are the precise language for asking: when does locally consistent data extend to a globally consistent picture? sheaf cohomology $H^n(X, \mathcal{F})$ measures the obstruction. $H^0$ is global sections. $H^1 \neq 0$ means locally consistent data that cannot be globally assembled — a topological contradiction encoded algebraically
the passage from a topological space to a site (a category equipped with a coverage axiom) and then to a topos (the category of sheaves on a site) generalizes topology beyond point-set foundations. every topos has an internal logic; every topological space is a special case
branches
point-set topology (general topology) — foundations: separation axioms (Hausdorff, normal, regular), compactness, connectedness, metrization theorems. the infrastructure on which all other branches rest
algebraic topology — homotopy groups, homology, cohomology, spectral sequences, K-theory. turns topological questions into algebraic computations
differential topology — smooth manifolds, tangent bundles, Morse theory, cobordism. topology with a smooth structure, the arena of relativity and quantum mechanics
geometric topology — knots, 3-manifolds, geometric structures (hyperbolic, spherical, Euclidean). the Poincaré conjecture (proved by Perelman, 2003) and geometrization belong here
topology in the cybergraph
knowledge topology is the shape of knowledge as revealed by graph structure. the Laplacian $L = D - A$ encodes topology algebraically; its spectral gap $\lambda_2$ (Fiedler value) measures how well-connected — how topologically robust — the knowledge is
the tri-kernel fixed point, the focus distribution $\pi^*$, is a sheaf-theoretic object: the unique global section consistent with every local diffusion, spring, and heat constraint. sheaf cohomology of the cybergraph measures contradictions in the knowledge structure that linking cannot resolve without topological change
the Seven Bridges of Koenigsberg — Euler's 1736 problem that founded graph theory — was the first topological argument: the question had no answer that depended on distances or shapes, only on connectivity. the cybergraph inherits this tradition
see also: category theory, algebra, sheaf, knowledge topology, Laplacian, spectral gap, homology, differential equations