Residue theorem and contour integration

A function that's perfectly well-behaved (holomorphic) everywhere except at a handful of isolated points can still be integrated around a closed path that avoids those points, and the residue theorem says the answer only depends on what happens at the enclosed singularities: ∮f(z)dz = 2πi · Σ(residues inside the contour). Everywhere the function is holomorphic contributes nothing at all — deforming the contour through a holomorphic region never changes the integral, which is exactly why the singularities are the only thing that matters.

The residue itself is a single complex number extracted from the function's Laurent series at that point — for the common case of a simple pole, it's just the limit, as z approaches z₀, of (z − z₀)f(z): the coefficient that survives when you cancel the pole and take the limit. Higher-order poles need a derivative-based generalisation of that formula, which is exactly the kind of detail that matters for computing a residue correctly but has no bearing on whether a pole counts at all: any pole strictly inside the contour contributes its residue, full stop, regardless of its order.

This machinery turns real integrals that resist ordinary calculus — many integrals of rational or trigonometric functions over the whole real line — into a two-step recipe: close the real-axis path with a large semicircular arc, show the arc's contribution vanishes as its radius grows, and read off the answer as 2πi times the sum of the residues enclosed. The real integral was never computed directly; it fell out as the by-product of a complex one.