This page assumes nothing. Every squiggle, letter, and word the parent note used is unpacked here, in the order you need them. If you can read a graph, you can finish this page.
Figure s01 — a dark chalkboard. A horizontal white line (the x-axis) and a vertical white line (the y-axis) cross at the centre. A yellow dot sits up and to the right; a blue dashed line drops from it to the x-axis (showing "go right x") and a pink dashed line runs left to the y-axis (showing "go up y"). The dot is labelled "point (x, y)."
Why we need this: the whole topic is about a curve — a set of points — and its steepness. Steepness is an up-vs-right comparison, so we must first agree on what "up" and "right" mean. That agreement is the plane.
Why the topic needs it: we are about to build the fraction hf(a+h)−f(a), which uses two input values, a and a+h. Both must be things f can eat. This quietly matters at the edge of a domain (e.g. x at x=0, where only h>0 keeps us legal — so only a one-sided approach is possible there).
Figure s02 — the curve y=f(x) climbs across the board. On the x-axis a yellow dot marks a and a blue dot marks a+h; a pink arrow between them is labelled "step h." Two dashed vertical lines rise from these inputs up to the curve, meeting it at heights labelled f(a) (yellow) and f(a+h) (blue).
Why the topic needs it: you cannot measure steepness at a single point (nothing to compare against). hmanufactures a second point so we have something to compare. Later we shrink h toward 0 — that is the heart of the whole chapter.
This is the most important idea to nail before anything else.
Figure s03 — a straight white line slopes upward. Two yellow dots sit on it; a blue horizontal segment between them (labelled "run — sideways") and a pink vertical segment (labelled "rise — up") form a right triangle under the line. Caption text reads "slope = rise / run."
Why the topic needs it: "slope of the tangent" is the answer the derivative gives. Slope is rise/run — this exact fraction reappears as the difference quotient in Section 6.
Figure s05 — the curve y=f(x) again. A yellow dot at (a,f(a)) and a blue dot at (a+h,f(a+h)) are joined by a dashed secant line. A blue horizontal segment underneath is labelled "run = h" and a pink vertical segment on the right is labelled "rise = f(a+h) − f(a)." Caption: "avg rate = rise / run."
Why the topic needs it: the derivative is the limit of this quantity. It is the raw material we are about to refine.
Figure s04 — the curve with a yellow dot fixed at a. Three faint-to-bright blue secant lines pass through a and second points that creep closer to a; as they do, the secants rotate. A bright pink line — the tangent — is the line they settle onto. Labels: "secants (blue)" and "tangent (pink)."
Why the topic needs it: the secant slope is computable (two real points, honest division). The tangent slope is what we want but can't compute directly. The bridge between them is the limit.
Before writing the fraction, be crystal-clear about the two points whose slope we are taking:
Why the topic needs it: this fraction is the object the limit acts on. Master its two pieces (rise on top, run h on bottom) and the derivative formula is no longer scary.
Why the topic needs it: the limit is the machine that dodges 00. It lets us reach the instant (the single point) legally — but only when the two sides agree. Everything rigorous about "→0" and one-sided limits lives in Limits — formal definition.
Why the topic needs it: it is the final packaged answer — one number that is both the tangent's slope and the instantaneous rate. Letting a roam over all inputs upgrades f′(a) into a whole new function, the Derivative as a function.
Read this map bottom-up as a build order. The coordinate plane is the stage; on it lives a functionf(x) with its domain of legal inputs. Choosing a spot a and a step h (with a+h still legal) gives us two points, whose slope — the same rise-over-run that measures steepness — becomes the average rate of change and the secant line. Feeding the difference quotient into a two-sided limit (which needs fcontinuous at a) finally produces the derivativef′(a).