Real Analysis 48 | Riemann Integral - Partitions [dark version]

Q: Why does the integral use “oriented” area, and how does that affect the calculation?

Oriented area assigns a sign based on position relative to the x-axis. Portions of the graph above the x-axis contribute positively, while portions below contribute negatively. In the rectangle approximation, this sign is carried by the rectangle heights: if f is negative on a subinterval, the corresponding rectangle area Cj (x_j − x_{j-1}) becomes negative, and summing preserves the overall signed area.

Q: What exactly is a partition of an interval [A, B]?

A partition is a finite ordered set of points x0, x1, …, xn such that x0 = A, xn = B, and x0 < x1 < … < xn. These points divide [A, B] into subintervals [x_{j-1}, x_j]. The partition is the backbone of the rectangle method because it determines the rectangle widths.

Q: What makes a function a step function in this setup?

A step function is piecewise constant on [A, B] with only finitely many jumps. Once the partition points are chosen to match the jump locations, the function is constant on each open subinterval (x_{j-1}, x_j), taking a value Cj there. The values at the jump points themselves are irrelevant for the integral because single points contribute no area.

Q: How is the Riemann integral defined for step functions?

For step functions, the integral is computed directly as a finite sum of rectangle areas. On each subinterval (x_{j-1}, x_j), the rectangle has height Cj and width (x_j − x_{j-1}). The integral equals Σ_{j=1}^n Cj (x_j − x_{j-1}). No limiting process is needed because the function already matches the rectangle structure exactly.

Q: Why must the step-function integral be “well-defined”?

A step function can often be represented using different partitions or different choices of constants Cj (for instance, by refining the partition or using redundant points). Even if the partition changes, the computed sum must stay the same for the integral to be a legitimate definition. The next step is proving that any valid representation yields the same total Σ_{j=1}^n Cj (x_j − x_{j-1}).

TL;DR

Riemann integration computes signed (oriented) area: above the x-axis counts positively and below counts negatively.

Briefing Cornell Notes

Briefing

Riemann integration is introduced as the practical way to compute the signed (or “orientated”) area between a function’s graph and the x-axis. The sign matters: regions above the x-axis contribute positively, while regions below contribute negatively. To turn that geometric idea into a calculation, the method approximates the area using rectangles built from a partition of the interval [A, B]. Each rectangle’s width comes from the spacing between consecutive partition points, and its height comes from the function value chosen within that subinterval. Summing the rectangle areas yields an approximation to the integral, and the full Riemann integral is obtained by taking a limit as the partition is refined.

The core workflow starts with a partition: a finite, ordered set of points x0, x1, …, xn with x0 = A and xn = B, where the points satisfy x0 < x1 < … < xn. This partition breaks [A, B] into subintervals [x_{j-1}, x_j]. For general functions, the integral is defined via a limiting process, but the discussion first restricts attention to step functions to make the idea concrete.

A step function is described as a piecewise constant function on [A, B]. Its graph consists of horizontal segments, with only finitely many jump discontinuities. Crucially, the function’s values at the jump points themselves do not affect the integral, because those points have no area contribution. Formally, once a partition is chosen so that its points align with the jump locations, the function is constant—equal to some constant Cj—throughout each open subinterval (x_{j-1}, x_j). Even if two neighboring subintervals share the same constant value, that causes no issue; what matters is that each subinterval has a constant height.

For step functions, the integral becomes a finite sum rather than a limit. Since each subinterval corresponds to one rectangle, the signed area is computed by adding rectangle areas: height Cj times width (x_j − x_{j-1}), summed over j = 1 to n. The remaining conceptual task is to justify that this definition is well-defined: although there may be many ways to choose partitions and constants Cj that represent the same step function, the resulting sum must be the same. That well-definedness question is flagged as the key point to resolve next, before moving on to the general Riemann integral via limits.

Finally, the discussion situates Riemann integration within a broader landscape by contrasting it with the Lebesgue integral. Both target the same signed area idea, but the Lebesgue integral is described as more general in what kinds of functions it can handle, making Riemann integration a natural starting point for building intuition and formal definitions.

Cornell Notes

Riemann integration computes the signed area between a function f and the x-axis on an interval [A, B]. The method begins by splitting [A, B] using a partition x0 < x1 < … < xn with x0 = A and xn = B, then approximating the area with rectangles whose widths are (x_j − x_{j-1}) and whose heights come from function values on each subinterval. For step functions—functions that are constant on each (x_{j-1}, x_j)—the integral reduces to a finite sum: Σ_{j=1}^n Cj (x_j − x_{j-1}). A key next step is proving this sum is well-defined even though different partitions or representations might be used for the same step function. This sets up the later limit process needed for general Riemann integrals.

Why does the integral use “oriented” area, and how does that affect the calculation?

Oriented area assigns a sign based on position relative to the x-axis. Portions of the graph above the x-axis contribute positively, while portions below contribute negatively. In the rectangle approximation, this sign is carried by the rectangle heights: if f is negative on a subinterval, the corresponding rectangle area Cj (x_j − x_{j-1}) becomes negative, and summing preserves the overall signed area.

What exactly is a partition of an interval [A, B]?

A partition is a finite ordered set of points x0, x1, …, xn such that x0 = A, xn = B, and x0 < x1 < … < xn. These points divide [A, B] into subintervals [x_{j-1}, x_j]. The partition is the backbone of the rectangle method because it determines the rectangle widths.

What makes a function a step function in this setup?

A step function is piecewise constant on [A, B] with only finitely many jumps. Once the partition points are chosen to match the jump locations, the function is constant on each open subinterval (x_{j-1}, x_j), taking a value Cj there. The values at the jump points themselves are irrelevant for the integral because single points contribute no area.

How is the Riemann integral defined for step functions?

For step functions, the integral is computed directly as a finite sum of rectangle areas. On each subinterval (x_{j-1}, x_j), the rectangle has height Cj and width (x_j − x_{j-1}). The integral equals Σ_{j=1}^n Cj (x_j − x_{j-1}). No limiting process is needed because the function already matches the rectangle structure exactly.

Why must the step-function integral be “well-defined”?

A step function can often be represented using different partitions or different choices of constants Cj (for instance, by refining the partition or using redundant points). Even if the partition changes, the computed sum must stay the same for the integral to be a legitimate definition. The next step is proving that any valid representation yields the same total Σ_{j=1}^n Cj (x_j − x_{j-1}).

Review Questions

Given a partition x0 < x1 < … < xn of [A, B], write the formula for the width of the j-th rectangle and explain what determines its height.
Explain why the value of a step function at a jump point does not affect the integral.
For a step function with constants C1, …, Cn on (x0, x1), …, (x_{n-1}, x_n), compute the integral as a finite sum.

Key Points

1
Riemann integration computes signed (oriented) area: above the x-axis counts positively and below counts negatively.
2
A partition of [A, B] is a finite ordered set x0 < x1 < … < xn with x0 = A and xn = B.
3
Rectangle approximations use widths (x_j − x_{j-1}) from the partition and heights from function values on each subinterval.
4
For step functions, the integral is a finite sum Σ_{j=1}^n Cj (x_j − x_{j-1}) because the graph is already rectangle-like.
5
Values at jump points of a step function do not affect the integral since single points contribute zero area.
6
The definition must be shown well-defined: different partitions/representations of the same step function must produce the same sum.
7
Lebesgue integration is mentioned as a more general alternative that still captures the same signed-area idea for a wider class of functions.

Highlights

Signed area is built into the calculation: negative function values produce negative rectangle contributions.

A partition is the structural input—its point spacing directly sets rectangle widths.

Step functions turn the integral into a clean finite sum of rectangle areas.

Jump-point values don’t matter because they have no area contribution.

The next conceptual hurdle is proving the step-function sum is independent of how the partition is chosen.