In the following section, Set Theory is presupposed with the Euclidean norm \(||\cdot||\).

Definition: A family of sets \(\mathbb{Y} \subseteq \mathcal{P}(X)\) is called topology on \(X \subseteq R\) if every intersection and union of sets of \(\mathbb{Y}\) belongs apart from \(\emptyset\) and \(X\) to \(\mathbb{Y}\). The pair \((X, \mathbb{Y})\) is called topological space. If \(\mathbb{Y} = \mathcal{P}(X)\), the topology is called discrete. A set \(B \subseteq \mathbb{Y}\) is called a base of \(\mathbb{Y}\) if every set of \(\mathbb{Y}\) can be written as union of any number of sets of \(B\). Every irreflexive relation \(N \subseteq {A}^{2}\) founds a NR in \(A \subseteq X\) for the underlying set \(X\).\(\triangle\)

Definition: If \((a, b) \in N\), \(a\) is called neighbour of or neighbouring to \(b\). Particularly, an element \(x \in A \subseteq X\) is called neighbour of an element \(y \in A\), where \(x \ne y\) if for all \(z \in X\) and a mapping \(d: {X}^{2} \rightarrow \mathbb{R}_{\ge 0}\) holds: (1) \(d(x, y) \le \text{max }\{\text{min }\{d(x, z), d(z, x)\}, \text{min }\{d(y, z), d(z, y)\}\}\) and (2) \(d(z, z) = 0\). Here \(d\) is called neighbourhood metric. Let \(P = R \cup V\) be the set of all points partitioned into actual points \(R\) and virtual points \(V\) for \(R, V \ne \emptyset = R \cap V\). When \(R\) or \(V\) is clear from context, it can be omitted.\(\triangle\)

Definition: The set \(A^{\prime} := R \setminus A\), where \(A \subseteq R\), is called complement of \(A\) in \(R\). \(A^{\prime}\) can be called the exterior of \(A\). All points of \(V \; (A)\) that have a neighbour in \(R \; (A^{\prime} \cup V)\) form the (inner) boundary \(\partial V \; (\partial A)\) of \(V \; (A)\). Here \({}^{\prime}\) takes precedence over \(\partial\). When \(\partial\) is applied successively beyond that, the argument is assumed to be without complement. The set \(A° := A \setminus \partial A\) is called the interior of \(A\). If \(\partial A \subseteq A\) is increased by the condition min \(\{d(x, y) : x \in A°, y \in A^{\prime}\} = \tilde{\nu}\), let \(A^{\ll} := A \setminus \partial A.\triangle\)

Definition: A set \(S \subseteq R \; (V)\) is said to be connected if there is for every partition of \(S\) into \(Y \cup Z\) such that \(Y, Z \ne \emptyset = Y \cap Z\): \(\partial Y^{\prime} \cap \partial Z \ne \emptyset \ne \partial Z^{\prime} \cap \partial Y\). \(S \subseteq R\) is moreover said to be simply connected if holds: Both \(\partial Y^{\prime} \cap \partial Z \cup \partial Z^{\prime} \cap \partial Y\) for every partition into connected \(Y\) and \(Z\) and \(S^{\prime} \cup (\partial)V\) for \( S^{\prime}\) as complement of \(S\) in \(R\) are connected for a connected (\(\partial)V\). Let \(P\) and \(R\) be simply connected. Every \(U \subseteq R\) is called neighbourhood of \(x \in R\) if \(x \in U°.\triangle\)

Definition: If \(\emptyset \ne \mathbb{D} \subseteq (X, \mathbb{Y})\) holds, a connected \(\mathbb{D}\) is called domain. The set of \(h\)-\(S \subseteq \mathbb{R}^{m}\) for \(m \in \mathbb{N}^{*}\) is \(n\)-dimensional, where \(m \ge n \in \mathbb{N}^{*}\), if and only if it contains at least one \(n\)-cube with edge length \(h \in \mathbb{R}_{>0}\) and maximum \(n\). Let \({}^1\dot{\mathbb{R}}^n\) the unit ball with the special case unit disk \({}^1\dot{\mathbb{R}}^2\). Midpoints \(a\) of \(n\)-balls and \(n\)-cubes may be denoted in brackets as (\(a\)) after those. A set has dimension \({}_{\tilde{\iota}}n\) if its elements consist of \(n\) maximal cubes of edge length \(\iota.\triangle\)

Examples: The base for \(\mathbb{N}, \mathbb{Z}, \mathbb{A}_\mathbb{R}, \mathbb{A}_\mathbb{C}, \mathbb{R}\) and \(\mathbb{C}\) is precisely each related discrete topology. The boundary of every \(n\)-ball with \(n \ge 2\) is only connected, itself it is simply connected. For \(n = 1\) both are not connected. For \(n \ge 2\) and \(r \in \mathbb{R}_{>0}\), the \(n\)-Torus \({}^r\mathbb{T}^n := \left(\partial{}^r\dot{\mathbb{R}}\right)^n\) is only connected.

Theorem: For \(n \in \mathbb{N}_{\ge 2}\), no finite decomposition of an \(n\)-ball can be reassembled giving an \(n\)-cube, since finitely many convex boundaries cannot have the same order of concave counterparts.\(\square\)

Theorem: Traversing up to nine cells of a tic-tac-toe grid renders the singleton the only connected set with fixed point property (invalid hairy ball theorem).\(\square\)

Definition: A function between two topological spaces is said to be continuous if for every point that can be mapped holds: for every neighbourhood of the image of this point there is a neighbourhood of the point whose image lies completely in the neighbourhood of the image of this point.\(\triangle\)

Remark: The suggestive terms compactness and countability (possibly misleading) are not used since they are not given for infinite sets like \(\mathbb{R}\) and \(\mathbb{N}\) resp. For all contracting deformations, points are to be removed from the target set as the contraction specifies. Only then the (generalised) Poincaré conjecture holds.

Remark: The neighbouring boundary points of the conventional closed [0, 1] and the conventional open ]0, 1[ especially have not the Hausdorff property1see Querenburg, Boto von: Mengentheoretische Topologie; 3., neu bearb. u. erw. Aufl.; 2001; Springer; Berlin, p. 83. So not every metric space can be a Hausdorff space or normal and (pre-) regular spaces are limited. The spaces \(\mathbb{C}^{n}\) and \(\mathbb{R}^{n}\) with \(n \in {}^{\omega }\mathbb{N}^{*}\) have therefore only the Fréchet topology2cf. Kowalsky, Hans-Joachim: Topologische Räume; 1. Aufl.; 1961; Birkhäuser; Basel, p. 62 ff.. The situation is, however, different in partially imprecise conventional mathematics.

© 2013-2019 by Boris Haase



1 see Querenburg, Boto von: Mengentheoretische Topologie; 3., neu bearb. u. erw. Aufl.; 2001; Springer; Berlin, p. 83
2 cf. Kowalsky, Hans-Joachim: Topologische Räume; 1. Aufl.; 1961; Birkhäuser; Basel, p. 62 ff.