Before you can read a single line of the parent note, you need to own every symbol it throws at you. We build them one at a time, each from the picture it names.
Picture a ruler laid flat. The left tick is x=a, the right tick is x=b. Everything we do lives on this ruler.
Figure 1 — the interval on a ruler. A plain horizontal ruler for the interval [a,b]: the left tick marked x=a and the right tick marked x=b in plum, with the shaded band between them (teal) showing "every position we care about lives here." No curve or heights yet — those symbols arrive in Sections 2 and 3.
Why the topic needs it: a BVP is about what happens between two ends. Those two ends are exactly x=a and x=b.
Figure 2 — the pinned-end picture. Now that y, α and β are defined, a teal curve y(x) is stretched over the ruler [a,b], with orange dots nailing its left end (y(a)=α) and right end (y(b)=β); the two ends are marked by plum dashed lines at x=a and x=b.
Why the topic needs it: the whole difference between an IVP and a BVP is where these numbers live. Same α at the left, but a BVP pins β at the far end while an IVP pins the slopeγ at the same end.
The prime mark ′ just means "the slope of." Read y′ as "y-prime."
Figure 3 — the derivative as a tangent's steepness. A teal curve with an orange tangent arrow touching it at one plum point; a small plum dotted "rise over run" triangle shows that y′ is exactly that rise divided by that run.
Why the topic needs it: shooting guessesy′(a) (the launch angle). If you have never pictured a slope as the steepness of a tangent arrow, the phrase "guess the missing slope" is meaningless.
Recall What does the prime in
y′ mean, in one word?
Slope ::: the steepness of the tangent arrow to the curve.
Why the topic needs it: every BVP the parent studies is a second-order ODE — the governing law is written in terms of y′′. The finite-difference method's core trick is turning y′′ into arithmetic on three dots.
Why the topic needs it: the parent's Taylor expansion contains y′′′ and y′′′′ in its remainder terms. Those terms are what decide the order of accuracy (O(h2)), so you must know they are simply "slopes-of-slopes-of-slopes."
Recall What does the notation
y(3) mean?
The third derivative ::: take "the slope of" three times; same as y′′′.
Think of it as a law of physics saying: at every point, the amount you bend is dictated by where you are and how you're leaning. The rope is not free — it must satisfy this rule at all of its infinitely many points at once.
Why the topic needs it: this is the problem. Both methods exist only to find the rope obeying this rule and the two end conditions.
Why the topic needs it: the parent splits its whole treatment on this word — the "2-shot shortcut" and the tridiagonal setup both require the linear form.
Recall Spot the nonlinear one:
y′′=xy′+y vs y′′=yy′+x.
The second ::: yy′ multiplies two unknowns together, so it is nonlinear; the first is linear.
Figure 4 — the rope becomes a beaded necklace. The faint teal true rope with orange beads y0,…,yN sitting at ticks x0,…,xN; the two plum end-beads are the fixed boundary values, the orange interior beads are the unknowns, and a labelled bracket shows the step h.
Why the topic needs it: every finite-difference formula is written with yi−1,yi,yi+1 — a bead and its two neighbours. Subscripts are the language of the method.
Why the topic needs it: the central-difference stencils 2hyi+1−yi−1 and h2yi+1−2yi+yi−1 are derived from Taylor. Without Taylor they look like magic.
Figure 5 — order of accuracy as a straight line on log–log paper. Error plotted against step size h with both axes on a log scale: a first-order O(h) method (teal) falls along a line of slope 1, while a second-order O(h2) method (orange) falls twice as steeply, slope 2 — halving h shifts the orange dot down by a factor of 4.
Why the topic needs it: "second-order" is the parent's headline accuracy claim, and the reason it warns against the cheaper O(h) forward difference. See Truncation error and order of accuracy.
Recall If a scheme is
O(h2) and you cut h in half, by what factor does the error fall?
By 4 ::: because (1/2)2=1/4.
Figure 6 — the secant method finding the root of ϕ. The miss-curve ϕ(s) plotted against the guessed slope s; two orange tried points are joined by a plum secant line, and where that line crosses the horizontal axis (teal) is the next, better guess sn+1 — repeat and you march onto the root ϕ=0.
Why the topic needs it: shooting is root finding stacked on an IVP solver — this vocabulary is half the method.
Figure 7 — the tridiagonal system Ay=d laid out. The matrix A shown with only three shaded stripes (orange diagonal, teal off-diagonals) and zeros elsewhere; the vector y of unknown interior heights y1,…,yN−1; and the right-hand vector d whose first and last entries carry the plum boundary values α,β.
Why the topic needs it: shooting reuses marching (RK4); finite difference produces this tridiagonal Ay=d. Knowing what A, y and d hold is essential to reading the parent's setup.
