the Cauchy kernel
/ koh-SHEE /
Inside every Cauchy integral sits the same little gadget: the expression 1 / (z - w), the Cauchy kernel. It is the weighting function that, when you integrate f against it around a loop, reproduces f at a point. Understanding the kernel is understanding why the whole formula works — it is a concentrated spike that samples f exactly where you want it.
Concretely the Cauchy kernel is 1 / (w - z_0) (or, written the other way, the differential form dz / (z - z_0)) appearing in f(z_0) = (1 / (2 pi i)) times the integral over gamma of f(z) / (z - z_0) dz. Its key features are simple but decisive: it is holomorphic in z except for a single simple pole at z = z_0, and integrating it alone around a loop enclosing z_0 gives exactly 2 pi i (and 0 if z_0 is outside). That clean '2 pi i or 0' is the winding number doing the bookkeeping, and it is what isolates the value at z_0 out of the whole boundary integral.
The kernel is the seed of the entire consequence tree. Differentiate it n times in z_0 and you get the kernel 1 / (z - z_0)^(n+1) of the generalized formula; expand it as a geometric series and you get Taylor coefficients; its single pole is what makes the residue calculus tick. In more advanced settings the same kernel reappears as the Cauchy transform and the Hilbert transform. It is, in a sense, the most important single fraction in complex analysis.
Integrating the bare kernel: the integral over the unit circle of dz / (z - 0) equals 2 pi i, while the integral over the same circle of dz / (z - 2) equals 0 because the pole at 2 lies outside.
The kernel's loop integral is 2 pi i if the pole is enclosed, 0 otherwise.
The 2 pi i is not magic decoration — it is the winding number of the contour about z_0 times 2 pi i; for a curve that loops twice the integral doubles to 4 pi i, so orientation and winding genuinely matter.