Uniform vs Pointwise Convergence
A sequence of functions can converge in two quite different senses, and the gap between them is where a lot of analysis lives. Pointwise convergence is the modest claim that at each fixed $x$ the numbers $f_n(x)$ settle to a limit $f(x)$, each point minding its own business. Uniform convergence is the demanding claim that the entire graph settles together, that the worst-case vertical gap between $f_n$ and $f$, the sup-norm $\lVert f_n - f\rVert_\infty$, shrinks to zero. The first does not imply the second, and the playground shows why. Sweep the index $n$ and watch the family evolve: the limit is drawn as the dashed curve, $f_n$ as the solid one, and the red bar marks the single worst point, the height of the sup-norm. For $x^n$ on $[0,1]$ every point below one tends to zero, yet a stubborn shoulder near $x=1$ keeps the gap at one and the limit ends up discontinuous, a continuity that uniform convergence would have protected. The bump examples are starker still: a spike of fixed or even growing height slides toward the edge, so $f_n(x)\to 0$ everywhere while the sup-norm refuses to fall. Only the flattening ramp converges uniformly, its whole graph pressed down to zero. The bottom panel plots that sup-norm against $n$, the one number that tells uniform from merely pointwise, and the draggable point on top tracks $f_n(x_0)$ settling pointwise even when the sup-norm will not.
WHAT TO TRY
- Watch the sweep: for the ramp the whole curve flattens onto zero (uniform), but for the bumps a spike slides to the edge and the gap stays open.
- On $x^n$, see the limit turn discontinuous, a jump at $x=1$ that no continuous uniform limit could produce.
- On the tall bump, the spike grows without bound even as every fixed point tends to zero, so the sup-norm diverges.
- Drag the probe $x_0$: its value $f_n(x_0)$ always settles (pointwise), even when the red gap and the sup-norm do not.