the zeros of an analytic function
A zero of a function f is simply a point z_0 where f(z_0) = 0 — a place where the function crosses through nothing. For polynomials these are the roots, and a degree-n polynomial has exactly n of them counted with multiplicity (the fundamental theorem of algebra). For a general analytic function the zeros are richer and more structured, and understanding them unlocks the identity theorem, the argument principle, and the whole study of where complex functions vanish.
The key structural fact: if f is analytic and not identically zero, then near each of its zeros z_0 you can factor out a power of (z - z_0). Concretely, f(z) = (z - z_0)^m g(z) where m is a positive integer (the order of the zero) and g is analytic with g(z_0) not equal to 0. You see this directly from the Taylor series: expand f around z_0, and since f(z_0) = 0 the constant term is gone; m is the index of the first nonzero coefficient, so f starts a_m (z - z_0)^m + higher terms, and pulling out (z - z_0)^m leaves a series that does not vanish at z_0. This factorization is the analytic mirror of factoring a polynomial.
From the factorization flow two profound consequences. First, the zeros of a nonzero analytic function are isolated: each zero sits alone with a small disk around it containing no other zero, because the nonvanishing factor g keeps f away from 0 nearby. Second, because zeros are isolated, they cannot accumulate inside the domain — if the zeros piled up toward an interior point, the function would be forced to be identically zero. This is exactly what powers the identity theorem and makes analytic functions astonishingly rigid: their zero sets are sparse, discrete, and informative.
f(z) = sin z is entire and vanishes exactly at z = n pi for integers n — an infinite but isolated set, no two zeros within distance pi of each other. Near z = 0, sin z = z - z^3/6 + ... = z (1 - z^2/6 + ...), so the zero at 0 has order 1 with nonvanishing factor g(z) = 1 - z^2/6 + ...
Even infinitely many zeros stay isolated and discrete — they only fail to be isolated where the function is identically zero.
Isolation of zeros requires the function to be analytic and not identically zero. A real smooth (but non-analytic) function can vanish on a whole interval; no nonzero analytic function can, which is precisely the gap between smooth and analytic.