Dirichlet conditions for convergence: Foundations

Before you can read the parent note, you must be able to read every squiggle in it without hesitation. This page defines each one from nothing — plain words, then a picture, then why the topic needs it. Nothing here assumes you've met calculus notation before.

0. The starting picture: a repeating shape

Everything begins with a function $f (x)$ . Think of $x$ as a position along a horizontal line, and $f (x)$ as the height of a curve above that position. Feed in a position, get back a height.

Figure — Dirichlet conditions for convergence

Why the topic needs it: the whole game is "can we rebuild this height-curve out of waves?" So we must first agree what a curve even is, and that it gives one height per position (condition 1 will demand this).

1. Periodic, and the period $2 L$

The parent writes "period $2 L$ ". Here is what that means.

Why $2 L$ and not just " $P$ "? Because the waves we'll use, $cos \frac{nπ x}{L}$ and $sin \frac{nπ x}{L}$ , are built to repeat exactly over width $2 L$ . Writing the period as $2 L$ makes the $L$ inside those waves line up cleanly — one full wave of the slowest sine fits one period. Look at the figure: over one width $2 L$ , the shape is copied left and right forever.

2. The interval $- L < x < L$ and endpoints

The parent keeps saying " $- π < x < π$ " or " $(- L, L)$ ". This is just naming the one period we inspect: positions from $- L$ up to $+ L$ . The centre $0$ sits in the middle; $- L$ and $+ L$ are the two edges where one copy ends and the next begins.

Why it matters: the endpoints $\pm L$ are exactly where a "wrap-around jump" can appear (Example 2 in the parent, $f (x) = x$ , jumps from $π$ down to $- π$ at the seam). You must picture the interval as a tile that gets repeated, so the right edge of one tile meets the left edge of the next.

3. Left and right limits: $f (x^{+})$ and $f (x^{-})$

This is the single most important piece of new notation in the parent, and it needs a picture.

Why the topic needs this above all else: at a jump, the curve has two honest heights at one position — the bottom of the step and the top of the step. Ordinary " $f (x_{0})$ " can't say which one. The pair $f (x_{0}^{-}), f (x_{0}^{+})$ names both, so the parent's midpoint formula

S (x_{0}) = \frac{f ( x _{0}^{+} ) + f ( x _{0}^{-} )}{2}

can talk about "halfway up the step". Look at the figure: the pink dot sits exactly between the blue (left) and yellow (right) levels.

4. Continuity vs a jump

Why the topic needs the distinction: condition 2 allows finitely many finite jumps but forbids infinite ones. The midpoint rule only makes sense at a finite jump (you can average two finite numbers; you cannot average $+ \infty$ ). See Piecewise smooth functions for the family of "curve made of a few smooth pieces glued at finite jumps".

5. Maxima and minima ("wiggles")

Why the topic needs it: condition 3 bans infinitely many wiggles. The parent's failing example $sin (1/ x)$ oscillates faster and faster as $x \to 0$ , cramming infinitely many peaks into a finite stretch — no finite pile of smooth waves can chase that, so convergence isn't guaranteed. Picture a spring compressed infinitely tight near a wall.

6. The integral sign $\int_{- L}^{L}$ and "absolutely integrable"

The scary symbol $\int$ just means "total signed area under the curve" between two positions.

Why $∣ f ∣$ and not just $f$ ? With plain $f$ , a curve could have huge positive area and huge negative area that cancel to a small number, hiding a blow-up. Taking $∣ f ∣$ folds everything up so nothing can cancel — this is exactly what condition 1 needs to be sure the coefficient integrals genuinely exist. See Fourier Series — coefficients via orthogonality for where those integrals come from.

7. The wave ingredients: $cos \frac{nπ x}{L}$ , $sin \frac{nπ x}{L}$

Why the topic needs them: these are the only pieces allowed in the stack. Rebuilding $f$ means choosing how much of each wave to add. The Dirichlet conditions decide whether this stack, taken to infinitely many waves, actually settles onto $f$ .

8. The coefficients $a_{0}, a_{n}, b_{n}$ and the " $\sim$ " warning

Why the topic needs it: the entire point of the parent note is upgrading " $\sim$ " to " $=$ " (or to the midpoint at jumps). See Convergence of series (pointwise vs uniform) for what "the sum reaches $f$ " precisely means.

9. The symbol $\sum$ and "partial sum" $S_{N}$

Why the topic needs it: the midpoint result is really a statement about the limit of $S_{N}$ as $N \to \infty$ . The Dirichlet kernel and Gibbs phenomenon are both stories about how these partial sums behave near a jump.

Prerequisite map

Equipment checklist

Recall Self-test: can you read every symbol? (hide and answer)

What does "single-valued function" mean in one sentence? ::: One position gives exactly one height — the curve never has two heights above the same $x$ . If the period is $2 L$ , what is $L$ and what equation states periodicity? ::: $L$ is the half-period, and $f (x + 2 L) = f (x)$ . In words, what is $f (x_{0}^{-})$ ? ::: The height the curve heads toward as you approach $x_{0}$ from the left (smaller $x$ ). In words, what is $f (x_{0}^{+})$ ? ::: The height the curve heads toward as you approach $x_{0}$ from the right (larger $x$ ). When is a point continuous, in terms of these limits? ::: When $f (x_{0}^{-}) = f (x_{0}^{+})$ . What is a finite jump vs an infinite discontinuity? ::: A finite jump has two different but finite approach-heights; an infinite discontinuity shoots off to $\pm \infty$ . What does one "wiggle" mean and which condition limits them? ::: A peak-and-valley (one max and one min); condition 3 allows only finitely many. What does $\int_{- L}^{L} g d x$ compute? ::: The total signed area under $g$ between $- L$ and $L$ . Why use $∣ f ∣$ for "absolutely integrable"? ::: So positive and negative areas can't cancel and hide a blow-up; we measure total silhouette area. What does the harmonic number $n$ control in $sin \frac{nπ x}{L}$ ? ::: How many wiggles fit in one period — larger $n$ is a faster wave. What does the " $\sim$ " sign warn you about? ::: Coefficients exist, but the sum may not equal $f$ — that's the separate convergence question. What is the partial sum $S_{N} (x)$ ? ::: The sum of only the first $N$ wave-terms; its limit as $N \to \infty$ is what we test.

Connections

4.7.04 Dirichlet conditions for convergence (Hinglish)
Fourier Series — coefficients via orthogonality
Dirichlet kernel
Gibbs phenomenon
Piecewise smooth functions
Convergence of series (pointwise vs uniform)
Half-range expansions