A subset of a topological space that every open cover admits a finite subcover; in metric spaces equivalently a set whose every sequence has a convergent subsequence (sequential compactness) or that is complete and totally bounded.

Principle

Compactness organizes infinitary coverings into finite information: global cover properties reduce to finitely many local pieces, enabling finiteness arguments for continuous images and optimization.

Demonstration

The closed interval [0,1] in the real line: any open cover of [0,1] has a finite subcover, and any sequence in [0,1] has a convergent subsequence whose limit lies in [0,1].

Misapplication

Treating boundedness alone as sufficient for compactness in infinite-dimensional normed spaces (e.g., the unit ball in an infinite-dimensional Banach space is bounded but not compact).

Consequence

Continuous functions on a compact set attain maxima and minima and are uniformly continuous; images of compact sets under continuous maps are compact, which yields many existence and approximation results.

Reversal

A noncompact set admits an open cover with no finite subcover; in metric terms it may contain a sequence with no convergent subsequence (examples: the whole real line or an unbounded discrete set).

Boundary

A topological property: definitions and equivalences depend on the ambient category (metric versus general topology). Sequential compactness equals cover compactness in metric spaces but not in arbitrary topological spaces; compactness does not imply finiteness of cardinality.

Semantic Tension

Compactness is often conflated with 'closed and bounded' because of Heine–Borel in Euclidean spaces; that characterization does not hold in general metric or topological spaces where total boundedness and completeness must be considered separately.

Synthesis

A compact set is a topological subset whose global covering behavior is reducible to finitely many local pieces, guaranteeing subsequence convergence in metric settings and enabling minimization and continuity-preserving mappings.