Proposition. Let (X,d)(X,d) be a metric space, let a∈Xa\in X, and let E⊂XE\subset X. Then

a∈int⁡(E)⟺∃ δ>0 such that B(a;δ)⊂E.a\in \operatorname{int}(E)\quad\Longleftrightarrow\quad \exists\,\delta>0 \text{ such that } B(a;\delta)\subset E.
Remarks

Context. This equivalence connects the "union of all open subsets" definition of to the ball-based definition of openness in metric spaces.

Proof sketch.

  • If a∈int⁡(E)a\in\operatorname{int}(E), then aa lies in some open set G⊂EG\subset E, so by openness there is a ball around aa contained in G⊂EG\subset E.
  • If B(a;δ)⊂EB(a;\delta)\subset E, then B(a;δ)B(a;\delta) is open and contained in EE, hence lies in the union defining int⁡(E)\operatorname{int}(E), so a∈int⁡(E)a\in\operatorname{int}(E).