Metric Spaces: Distance, Open Sets and Convergence Guide

A metric space is a set X together with a metric (distance function) d(x, y) satisfying four rules: distances are non-negative, d(x, y) = 0 exactly when x = y, distance is symmetric, and the triangle inequality d(x, z) ≤ d(x, y) + d(y, z) holds. This tiny axiom list captures everything calculus needs about "closeness", freed from coordinates.

Metric spaces matter because analysis happens far beyond the real line: in function spaces, in discrete settings, in geometry. Once convergence, continuity and compactness are defined through a metric, every theorem proved from the axioms works in all these worlds simultaneously.

Key Concepts

• Metric axioms: (1) d(x, y) ≥ 0; (2) d(x, y) = 0 ⟺ x = y; (3) d(x, y) = d(y, x); (4) d(x, z) ≤ d(x, y) + d(y, z).
• Euclidean metric: on ℝn, d(x, y) = √((x₁−y₁)² + … + (xn−yn)²), the ordinary straight-line distance.
• Discrete metric: d(x, y) = 0 if x = y and 1 otherwise; every two distinct points are equally far apart, and every set is open.
• Open ball: B(x, r) = {y : d(x, y) < r}, the set of points within distance r of x; open balls are the building blocks of open sets.
• Convergence: xn → x means d(xn, x) → 0 as real numbers; limits in a metric space are unique.

Formulas

• Convergence: xn → x ⟺ ∀ε > 0, ∃N: n ≥ N ⇒ d(xn, x) < ε.
• Taxicab metric on ℝ²: d((x₁,y₁), (x₂,y₂)) = |x₁−x₂| + |y₁−y₂|, the distance a taxi drives on a grid.
• Sup metric on bounded functions: d(f, g) = supx |f(x) − g(x)|; convergence in this metric is exactly uniform convergence.
• Continuity via metric: f is continuous at c ⟺ ∀ε > 0, ∃δ > 0: d(x, c) < δ ⇒ d(f(x), f(c)) < ε.

Worked Example

In ℝ² with the Euclidean metric, show that an = (1/n, 2/n) converges to (0, 0).

1. Compute the distance: d(an, (0,0)) = √((1/n − 0)² + (2/n − 0)²) = √(1/n² + 4/n²) = √5/n.
2. Let ε > 0 be given. We need √5/n < ε, i.e. n > √5/ε.
3. Choose N larger than √5/ε; then for all n ≥ N, d(an, (0,0)) = √5/n ≤ √5/N < ε.
4. Hence d(an, (0,0)) → 0, so an → (0,0) in the metric space ℝ².

Common Misconceptions

• "Every metric space looks like ℝn." The discrete metric gives a space where every singleton {x} equals the open ball B(x, 1/2) — nothing like Euclidean geometry.
• "Convergent sequences always have their limit nearby in the set." In ℚ with the usual metric, a sequence of rationals can converge to √2, which is not in ℚ; the space is not complete.
• "The triangle inequality is just obvious." It is an axiom that must be verified for each candidate metric; the proposed "distance" d(x,y) = (x−y)² on ℝ fails it.

Real-World Applications

• Function spaces in analysis and machine learning: measuring how close two functions or two data distributions are.
• Clustering algorithms: k-means and nearest-neighbour methods depend on a chosen metric on the data space.
• Fixed-point theorems: the Banach contraction principle in complete metric spaces guarantees solutions to equations and underlies many numerical methods.

A metric space is a set plus a distance obeying four simple axioms, and from those axioms flow convergence, open sets, continuity and completeness. The Euclidean, taxicab, discrete and sup metrics show how flexible the idea is — the same definition of "xn → x" serves geometry, analysis and data science, with the worked example showing the ε–N verification in action.