A function f(x) is continuous at a point x=a when three separate requirements all hold together: the function must actually be defined at a (and on a whole open interval around it), the limit limx→af(x) must exist as a single finite number, and that limiting value must be exactly equal to f(a) itself. If any one of these three fails — the value is missing, the two-sided limit does not exist, or the limit disagrees with the assigned value — the function is discontinuous at that point. An equivalent way to state the third condition is limh→0[f(a+h)−f(a)]=0, which says that as the input is nudged away from a by a shrinking amount h, the resulting change in output must also shrink to zero, ruling out any sudden jump right at a. In practice, checking continuity at a point almost always comes down to computing f(a) directly from its formula, computing the two-sided limit (often by simplifying a 0/0 form through factoring, rationalising, or a standard limit such as limx→0xsinx=1), and then comparing the two numbers.