This deck builds the topology of the real line from its order. It covers open sets as wiggle room and closed sets as complements that contain their limit points, the line and the empty set as the clopen pair, the union and intersection laws and why an infinite intersection of open sets can fail to be open, compactness via open covers together with the Heine-Borel theorem, sequential compactness, and the connectedness of intervals. It targets the misconceptions that closed means not open, that an infinite intersection of open sets is open, that boundedness alone forces compactness, and that closed means finite or bounded.
Subject: Foundations of Higher Mathematics · 110 slides · symbolic lesson
Open the interactive version of this deck · Homework for this lesson
Objectives
By the end of this deck you can:
1. Decide whether a subset of the real line is open, closed, both, or neither.
2. Compute limit points and the closure of a set, and prove a set is closed by its complement.
3. State the union and intersection laws for open and closed sets, and explain why infinite intersections of opens can fail to be open.
4. Define compactness by open covers, prove a set is not compact by exhibiting a bad cover, and apply the Heine-Borel theorem.
5. Connect completeness to Heine-Borel and recognize compactness as a finiteness-control tool, echoing propositional compactness.
Warm-up
Discussion prompt
Before we open Topology of R: Open, Closed & Compact: without looking back, what was the main idea of Infinite Series, and what could you do by the end of it that you could not do before?
Hint: One sentence for the idea, one for the skill. If the second one is blank, that is the part to revisit.
Answer:
A series is the limit of its partial sums, not a magical infinite addition. Builds the nth-term test, geometric and telescoping sums, the harmonic divergence proof, comparison / ratio / root tests, and absolute vs conditional convergence.
Concept
So far the real line has been an ordered field: you can add, multiply, and compare. Topology asks a new kind of question.
It asks which points are near which. On the line, near means small distance, and distance comes from the order and the absolute value.
\[ d(x,y) = \lvert x - y \rvert \]
Every idea in this deck, open, closed, and compact, is built from one primitive: the set of points within a chosen distance of a center.
Counterexample
Discussion prompt
Every idea in this deck, open, closed, and compact, is built from one primitive: the set of points within a chosen distance of a center.
That is stated as though it always holds. Do one of two things: produce a case where it fails, or say precisely what rules such a case out. "It just does" is not on the menu.
Hint: Hunt at the extremes first — zero, one, negative, empty, equal. If every extreme survives, the reason they survive is the proof.
Intuition
Picture standing at a point inside a set. The set is open near you if you can take a small step in either direction and still be inside.
A set is open when this is true at every one of its points: nobody is standing on an edge with nothing to their left or right.
This wiggle room is the whole idea. Everything else formalizes it or takes complements of it.
Analogy
Discussion prompt
Explain Open means wiggle room by analogy to something with no Foundations of Higher Mathematics in it at all — a queue, a recipe, a map, a bank balance, whatever fits. Then say where your analogy breaks.
Hint: An analogy that never breaks is not an analogy, it is the same idea wearing a hat. Find the seam — that is the part that is actually new.
Answer:
Picture standing at a point inside a set. The set is open near you if you can take a small step in either direction and still be inside.
Concept
The basic open piece of the line is the open interval: all points strictly between two endpoints, endpoints excluded.
\[ (a,b) = \{\, x \in \mathbb{R} : a < x < b \,\} \]
The strict inequalities are the point. No endpoint belongs, so from any member you still have room on both sides.
Explain it
Discussion prompt
Explain The open interval to a student a year behind you. No notation, no jargon they have not met — and it still has to be true.
Hint: If your explanation needs a symbol they have never seen, you are describing the notation rather than the idea.
Answer:
The basic open piece of the line is the open interval: all points strictly between two endpoints, endpoints excluded.
Concept
Fix a center and a positive radius. The neighborhood is the open interval of that radius around the center.
\[ N_\varepsilon(x) = (x - \varepsilon,\; x + \varepsilon) = \{\, y : \lvert y - x \rvert < \varepsilon \,\} \]
epsilon-neighborhood — For a point x and a positive radius epsilon, the set of all points within distance epsilon of x. It is the basic notion of closeness on the line and the building block of every open set.
Concept
A point of a set is an interior point when the set gives it wiggle room: some neighborhood of the point lies entirely inside the set.
\[ x \text{ is interior to } S \iff \exists\, \varepsilon > 0 \;\text{ with }\; N_\varepsilon(x) \subseteq S \]
Interior points are the ones standing safely inside, not on the rim. The set of all of them is called the interior.
Concept
A set is open exactly when every one of its points is an interior point. No point sits on an edge.
\[ S \text{ open} \iff \forall\, x \in S \; \exists\, \varepsilon > 0 : N_\varepsilon(x) \subseteq S \]
open set — A subset of the real line such that around each of its points there is an entire neighborhood contained in the set. Equivalently, a set equal to its own interior.
Definition probe
Sort into buckets
Every line below is part of the definition of epsilon-neighborhood or of open set — one or the other, never both. Put each where it belongs.
Ranking
Put in order
Put the moves of Worked example: every open interval is open into the order they have to happen.
Why: These are the moves of the worked example in the order it makes them, and each one is set up by the one before it. To prove a for-all statement about points of the set, fix one generic point and argue about it.
Worked example
Claim: the open interval from a to b is an open set. We must find wiggle room at an arbitrary member.
\[ \text{Show } (a,b) \text{ is open.} \]
Take an arbitrary point of the interval
Why: To prove a for-all statement about points of the set, fix one generic point and argue about it.
\[ \text{let } x \in (a,b), \text{ so } a < x < b \]
Choose the radius to be the distance to the nearer endpoint
Why: Both gaps are positive, so their minimum is positive. Using the smaller gap guarantees we do not spill past either end.
\[ \varepsilon = \min(x - a,\; b - x) > 0 \]
Check the neighborhood stays inside
Why: Any y within epsilon of x satisfies y greater than x minus epsilon at least a, and y less than x plus epsilon at most b, so y lies in the interval.
\[ \lvert y - x \rvert < \varepsilon \;\Rightarrow\; a \le x - \varepsilon < y < x + \varepsilon \le b \]
Verify with a concrete point
Why: Take the interval from 0 to 1 and the point one half. The nearer endpoint distance is one half, and the neighborhood of radius one half around one half is exactly the interval from 0 to 1, which sits inside. The general argument checks out.
\[ x = \tfrac12 \in (0,1),\quad \varepsilon = \tfrac12,\quad N_{1/2}(\tfrac12) = (0,1) \subseteq (0,1) \]
Picture it
Animation
Shows: Each line of the worked example "every open interval is open", appearing one at a time.
The same working the example does, in the order a tutor would write it.
Takeaway: Any y within epsilon of x satisfies y greater than x minus epsilon at least a, and y less than x plus epsilon at most b, so y lies in the interval.
Concept
The whole real line is open: around any point the neighborhood of radius one already sits inside, since everything is inside.
\[ \forall\, x \in \mathbb{R}: N_1(x) \subseteq \mathbb{R} \]
The empty set is open too, vacuously: there is no point in it that could fail the requirement. A for-all over nothing is automatically true.
\[ \emptyset \text{ is open (vacuously)} \]
Step zero
Discussion prompt
Worked example: a closed interval is not open — before any calculation: what is the plan? Name the moves in order, in plain English, without doing the arithmetic.
Hint: It starts with: Point at the endpoint 1
Answer:
Worked example
Claim: the closed interval from 0 to 1 is not an open set. To break openness we only need one bad point.
\[ \text{Show } [0,1] \text{ is not open.} \]
Point at the endpoint 1
Why: The endpoint belongs to the closed interval but has no room to its right, so it is the natural place openness fails.
\[ 1 \in [0,1] \]
Show every neighborhood of 1 escapes the set
Why: For any positive radius, the point one plus half that radius is greater than 1, hence outside the interval yet inside the neighborhood.
\[ \text{for any } \varepsilon > 0:\; 1 + \tfrac{\varepsilon}{2} \in N_\varepsilon(1),\quad 1 + \tfrac{\varepsilon}{2} \notin [0,1] \]
Verify with a numeric radius
Why: Take radius one tenth. The point one point zero five lies in the neighborhood of 1 but exceeds 1, so no neighborhood of the endpoint fits inside. The point 1 is not interior, so the set is not open.
\[ \varepsilon = 0.1:\; 1.05 \in (0.9,\,1.1),\quad 1.05 \notin [0,1] \]
Picture it
Animation
Shows: Each line of the worked example "a closed interval is not open", appearing one at a time.
The same working the example does, in the order a tutor would write it.
Takeaway: Take radius one tenth. The point one point zero five lies in the neighborhood of 1 but exceeds 1, so no neighborhood of the endpoint fits inside. The point 1 is not interior, so the set is not open.
Trap
Open and closed sound like opposites, so a set that is not open must be closed, and a set that is not closed must be open.
Under this reading, every set falls into exactly one of the two boxes, the way a light switch is on or off.
Open and closed are not opposites. Closed is defined by the complement being open, and a set can be both, or neither.
Name the two escapes from the dichotomy
Why: The whole line and the empty set are both open and closed at once. A half-open interval is neither open nor closed. So not-open does not imply closed, and the two words label independent properties.
\[ \mathbb{R},\ \emptyset:\ \text{both} \qquad [0,1):\ \text{neither} \]
Concept
A set is closed when its complement is open. This is the cleanest definition and the one to compute with.
\[ S \text{ closed} \iff \mathbb{R} \setminus S \text{ is open} \]
closed set — A subset of the real line whose complement is open. Equivalently, a set that contains all of its limit points, as we will prove shortly.
Intuition
Open sets have no edge points inside them. Closed sets are the reverse: any point the set gets arbitrarily close to is already a member.
If a sequence of members marches toward a target, a closed set already contains the target. Nothing leaks out at the boundary.
Concept
A point is a limit point of a set when every neighborhood of it, however small, contains a point of the set other than the point itself.
\[ p \text{ is a limit point of } S \iff \forall\, \varepsilon > 0:\; \big(N_\varepsilon(p) \setminus \{p\}\big) \cap S \neq \emptyset \]
The point being approached need not belong to the set. That is exactly what lets a set miss a limit point and so fail to be closed.
Concept
The opposite of a limit point is an isolated point: a member that has a neighborhood containing no other member of the set.
\[ p \in S \text{ is isolated} \iff \exists\, \varepsilon > 0:\; N_\varepsilon(p) \cap S = \{p\} \]
A point can be in the set yet not a limit point of it. Membership and being-approached are separate questions, and this distinction drives the next example.
Estimation
Predict first
Consider the set of reciprocals of the positive integers. We find its limit points and its closure.
Commit before you compute: what does Worked example: limit points and closure of the reciprocals come out to? A rough magnitude and the right form is enough — the point is to have something concrete to be wrong about.
Correct: Verify the witness at a concrete radius
Why: A prediction you can defend turns the computation into a check rather than a leap of faith — and an answer that contradicts it is caught on the spot. Take radius one hundredth. Solving one over n below one hundredth gives n greater than one hundred, so one over one hundred one is about zero point zero zero nine nine, which lies in the neighborhood of zero.
Worked example
Consider the set of reciprocals of the positive integers. We find its limit points and its closure.
\[ S = \left\{\, \tfrac{1}{n} : n \in \mathbb{N},\ n \ge 1 \,\right\} \]
Show each reciprocal is isolated, not a limit point
Why: Consecutive reciprocals have a positive gap, so a small enough neighborhood around one of them catches no other member. Isolated points contribute nothing to the limit points.
\[ \tfrac{1}{n} - \tfrac{1}{n+1} = \tfrac{1}{n(n+1)} > 0 \]
Show 0 is a limit point
Why: Zero is not in the set, but every neighborhood of it contains some reciprocal, because reciprocals shrink below any positive radius. This uses the Archimedean property.
\[ \forall\, \varepsilon > 0\ \exists\, n:\ 0 < \tfrac{1}{n} < \varepsilon \]
Assemble the closure
Why: The closure is the set together with all its limit points. The only limit point is zero, so we add it and nothing else.
\[ \overline{S} = S \cup \{0\} \]
Verify the witness at a concrete radius
Why: Take radius one hundredth. Solving one over n below one hundredth gives n greater than one hundred, so one over one hundred one is about zero point zero zero nine nine, which lies in the neighborhood of zero. Zero really is approached, confirming it as the sole limit point.
\[ \varepsilon = 0.01:\ n = 101 \Rightarrow \tfrac{1}{101} \approx 0.0099 < 0.01 \]
Picture it
Animation
Shows: Each line of the worked example "limit points and closure of the reciprocals", appearing one at a time.
The same working the example does, in the order a tutor would write it.
Takeaway: Take radius one hundredth. Solving one over n below one hundredth gives n greater than one hundred, so one over one hundred one is about zero point zero zero nine nine, which lies in the neighborhood of zero. Zero really is approached, confirming it as the sole limit point.
Concept
The closure of a set is the set together with all of its limit points. It is the smallest closed set that contains the original.
\[ \overline{S} = S \cup \{\, p : p \text{ is a limit point of } S \,\} \]
closure — The union of a set with all of its limit points. It is closed, it contains the set, and it is contained in every closed set containing the original, so it is the smallest such closed set.
Concept
The complement definition and the boundary picture agree. A set is closed if and only if it contains every one of its limit points.
\[ S \text{ closed} \iff S = \overline{S} \iff \big(p \text{ a limit point of } S \Rightarrow p \in S\big) \]
Read from left to right: a leaked limit point would be a point of the open complement approached by members, which no neighborhood inside the complement can allow. So no limit point leaks.
Missing information
Discussion prompt
Claim: the closed interval from a to b is a closed set. We use the cleanest tool, the complement.
What do you need to know — or decide — before the first line can be written? List everything the problem has to hand you.
Hint: Anything you would have to invent to get started is a thing the problem must supply.
Answer:
Everything not between the endpoints is either strictly below a or strictly above b. That is a union of two open rays.
Worked example
Claim: the closed interval from a to b is a closed set. We use the cleanest tool, the complement.
\[ \text{Show } [a,b] \text{ is closed.} \]
Write the complement
Why: Everything not between the endpoints is either strictly below a or strictly above b. That is a union of two open rays.
\[ \mathbb{R} \setminus [a,b] = (-\infty, a) \cup (b, \infty) \]
Each ray is open
Why: A point below a has room to its right up to a and unlimited room to its left, so a small neighborhood stays below a. The same holds symmetrically above b.
\[ x < a \Rightarrow N_{a-x}(x) \subseteq (-\infty, a) \]
The complement is a union of opens, hence open
Why: A union of open sets is open, a law proved shortly. So the complement of the interval is open.
\[ (-\infty, a) \cup (b, \infty) \text{ is open} \]
Verify by checking a boundary point stays put
Why: Take the interval from 0 to 1. Its complement is everything below 0 or above 1, and the point one half of one below 0, namely negative one half, sits in an open ray. Since the complement is open, the interval is closed, and the endpoints 0 and 1 are members, as a closed set should keep its edges.
\[ [0,1]:\ \mathbb{R} \setminus [0,1] = (-\infty,0) \cup (1,\infty)\ \text{open} \]
Picture it
Animation
Shows: Each line of the worked example "a closed interval is closed", appearing one at a time.
The same working the example does, in the order a tutor would write it.
Takeaway: Take the interval from 0 to 1. Its complement is everything below 0 or above 1, and the point one half of one below 0, namely negative one half, sits in an open ray. Since the complement is open, the interval is closed, and the endpoints 0 and 1 are members, as a closed set should keep its edges.
Concept
A boundary point of a set is one that is arbitrarily close to both the set and its complement: every neighborhood meets each side.
\[ p \in \partial S \iff \forall\, \varepsilon > 0:\ N_\varepsilon(p) \cap S \neq \emptyset \ \text{and}\ N_\varepsilon(p) \cap (\mathbb{R}\setminus S) \neq \emptyset \]
Open sets contain none of their boundary; closed sets contain all of it. This reframes the two definitions in one picture.
Concept
A set that is both open and closed is called clopen. On the real line only two subsets are clopen.
\[ \text{Clopen subsets of } \mathbb{R}:\quad \emptyset \ \text{and}\ \mathbb{R} \]
That there are no others is a deep fact: it is exactly the statement that the line is connected, cannot be split into two nonempty opens. We return to this at the end.
Fill the middle
Fill in the blanks
From Trap: a half-open interval must be one or the other — finish the line. Write what belongs on the right of the equals sign before you look.
[0,1) \stackrel} \text{open or closed}___}
Why: Producing the right-hand side unprompted is the difference between recognising this line and being able to use it. At 0 every neighborhood contains negative numbers outside the set, so 0 is not interior and the set is not open.
Trap
The interval that includes 0 but not 1 surely has to be either open or closed. Pick whichever the endpoints suggest.
\[ [0,1) \stackrel{?}{=} \text{open or closed} \]
It is neither. Openness fails at the included endpoint, and closedness fails at the excluded endpoint that is still a limit point.
Break both properties at once
Why: At 0 every neighborhood contains negative numbers outside the set, so 0 is not interior and the set is not open. Meanwhile 1 is a limit point not in the set, so the set omits a limit point and is not closed.
\[ 0 \text{ not interior} \Rightarrow \text{not open};\quad 1 \text{ a missed limit point} \Rightarrow \text{not closed} \]
Pattern
1. Test openness point by point
Why: Ask whether every member has a neighborhood inside the set. A single member on an edge, with the complement immediately beside it, kills openness.
2. Test closedness through the complement or limit points
Why: Either show the complement is open, or hunt for a limit point the set fails to contain. Finding one missed limit point kills closedness.
3. Combine the two independent verdicts
Why: Open and closed are separate yes-or-no questions, giving four outcomes: open only, closed only, both (only the empty set and the whole line), or neither.
\[ \text{four cases}: \ \text{open}, \ \text{closed}, \ \text{clopen}, \ \text{neither} \]
Notation
Annotate
From Pattern: classify any subset of the line — read this one piece at a time. What is each part doing?
On: \( \text{four cases}: \ \text{open}, \ \text{closed}, \ \text{clopen}, \ \text{neither} \)
Elimination
Eliminate the wrong options
Which best describes the interval that excludes 0 and includes 1?
3 of these 4 are wrong. Strike them one at a time, and say what rules each one out before you strike the next. The survivor is the answer.
Survives elimination: A
Why: At the included endpoint 1 every neighborhood contains points above 1 outside the set, so 1 is not interior and the set is not open. At the excluded endpoint 0, which is a limit point not in the set, closedness fails. So it is neither.
Check
Classify the half-open interval that excludes 0 and includes 1.
\[ (0,1] \]
Check your understanding
Which best describes the interval that excludes 0 and includes 1?
Answer: A
Why: At the included endpoint 1 every neighborhood contains points above 1 outside the set, so 1 is not interior and the set is not open. At the excluded endpoint 0, which is a limit point not in the set, closedness fails. So it is neither.
Concept
Take any collection of open sets, finite or infinite, and union them all. The result is open.
\[ \text{each } U_i \text{ open} \;\Rightarrow\; \bigcup_{i \in I} U_i \text{ open} \]
See why the proof needs no finiteness
Why: A point in the union lands in at least one of the pieces. That single piece already supplies a neighborhood inside itself, hence inside the whole union. One witness suffices, so the index set may be as large as you like.
Fill the middle
Fill in the blanks
From Finite intersections of opens are open — finish the line. Write what belongs on the right of the equals sign before you look.
\varepsilon = \min(\varepsilon_1, \ldots, \varepsilon_k) > 0
Why: Producing the right-hand side unprompted is the difference between recognising this line and being able to use it. A point in the intersection has a radius that works for each piece.
Concept
Intersect finitely many open sets and the result is still open. Finiteness now matters.
\[ U_1, \ldots, U_k \text{ open} \;\Rightarrow\; U_1 \cap \cdots \cap U_k \text{ open} \]
See where finiteness enters
Why: A point in the intersection has a radius that works for each piece. Taking the minimum of finitely many positive radii is still positive. With infinitely many pieces the minimum could drop to zero, and the argument breaks.
\[ \varepsilon = \min(\varepsilon_1, \ldots, \varepsilon_k) > 0 \]
Translation
\( U_1, \ldots, U_k \text{ open} \;\Rightarrow\; U_1 \cap \cdots \cap U_k \text{ open} \)
Draw it
Translate both ways. First write the expression above as a sentence with no symbols in it at all. Then cover it, and write your sentence back as notation. If the two versions disagree, the disagreement is the thing to fix.
Step zero
Discussion prompt
Worked example: an infinite intersection of opens collapses — before any calculation: what is the plan? Name the moves in order, in plain English, without doing the arithmetic.
Hint: It starts with: Zero survives every stage
Answer:
Worked example
Intersect the shrinking open intervals centered at 0, one for each positive integer. The intersection is a single point.
\[ \bigcap_{n \ge 1} \left(-\tfrac{1}{n},\ \tfrac{1}{n}\right) = \{0\} \]
Zero survives every stage
Why: Zero lies in each interval since each is centered at zero with positive radius. So zero is in the intersection.
\[ 0 \in \left(-\tfrac{1}{n},\ \tfrac{1}{n}\right) \text{ for all } n \]
Every nonzero point is eventually excluded
Why: Given a nonzero x, pick n larger than the reciprocal of its absolute value, so the interval at stage n is too narrow to hold x. Hence x is not in the intersection.
\[ x \neq 0 \Rightarrow \exists\, n:\ \tfrac{1}{n} < \lvert x \rvert \Rightarrow x \notin \left(-\tfrac{1}{n},\ \tfrac{1}{n}\right) \]
The result is not open
Why: The single point set has no wiggle room: no neighborhood of zero fits inside a lone point. So an infinite intersection of opens produced a non-open set.
\[ \{0\} \text{ is not open} \]
Verify with a concrete escapee
Why: Take x equal to zero point three. Choosing n equal to four gives the interval from negative one quarter to one quarter, which excludes zero point three. So only zero remains in the intersection, confirming the collapse to a single non-open point.
\[ x = 0.3,\ n = 4:\ 0.3 \notin (-0.25,\ 0.25) \]
Picture it
Animation
Shows: Each line of the worked example "an infinite intersection of opens collapses", appearing one at a time.
The same working the example does, in the order a tutor would write it.
Takeaway: Take x equal to zero point three. Choosing n equal to four gives the interval from negative one quarter to one quarter, which excludes zero point three. So only zero remains in the intersection, confirming the collapse to a single non-open point.
Fill the middle
Fill in the blanks
From Trap: infinite intersections of opens stay open — finish the line. Write what belongs on the right of the equals sign before you look.
\bigcap_\{0\} \ \text{not open}\left(-\tfrac1n,\tfrac1n\right) = ___
Why: Producing the right-hand side unprompted is the difference between recognising this line and being able to use it. The shrinking intervals around zero intersect to the single point zero, which is not open.
Trap
Unions of opens are open with no size limit, so surely intersections of opens are open too, no matter how many.
The reasoning copies the union rule and drops the word finite as if it were decorative.
Only finite intersections are guaranteed open. The word finite is load-bearing, not decoration.
Point at the counterexample and the reason
Why: The shrinking intervals around zero intersect to the single point zero, which is not open. The proof for finite intersections took a minimum of radii, safe for finitely many but not for infinitely many, where the minimum can be zero.
\[ \bigcap_{n \ge 1}\left(-\tfrac1n,\tfrac1n\right) = \{0\} \ \text{not open} \]
Break the constraint
Discussion prompt
The rule this trap just fixed:
Only finite intersections are guaranteed open. The word finite is load-bearing, not decoration.
Now break it on purpose. Build a case that violates it and follow the consequences until something visibly fails. Where does the failure first show up — and would you have noticed it if you had not been looking?
Hint: The dangerous rules are the ones whose violation still produces an answer. If yours fails loudly, try to find one that fails quietly.
Concept
Complementation swaps unions with intersections and open with closed. So the two open laws dualize into two closed laws.
\[ \mathbb{R} \setminus \bigcup_i A_i = \bigcap_i (\mathbb{R}\setminus A_i), \qquad \mathbb{R} \setminus \bigcap_i A_i = \bigcup_i (\mathbb{R}\setminus A_i) \]
Arbitrary intersections of closed sets are closed. Finite unions of closed sets are closed. The finiteness moves from intersection over to union.
\[ \bigcap_i C_i \text{ closed}; \qquad C_1 \cup \cdots \cup C_k \text{ closed} \]
Estimation
Predict first
Claim: the union of finitely many closed sets is closed. We prove it by pushing through complements.
Commit before you compute: what does Worked example: a finite union of closed sets is closed come out to? A rough magnitude and the right form is enough — the point is to have something concrete to be wrong about.
Correct: Verify on a concrete pair
Why: A prediction you can defend turns the computation into a check rather than a leap of faith — and an answer that contradicts it is caught on the spot. The union of the closed intervals from 0 to 1 and from 2 to 3 has complement equal to three open rays and gaps, all open, so the union is closed.
Worked example
Claim: the union of finitely many closed sets is closed. We prove it by pushing through complements.
\[ C_1, \ldots, C_k \text{ closed} \;\Rightarrow\; C_1 \cup \cdots \cup C_k \text{ closed} \]
Complement the union with De Morgan
Why: The complement of a union is the intersection of the complements. This turns the problem into one about open sets.
\[ \mathbb{R} \setminus \bigcup_{j=1}^{k} C_j = \bigcap_{j=1}^{k} (\mathbb{R} \setminus C_j) \]
Each complement is open
Why: Each set is closed, which is exactly the statement that its complement is open. So we are intersecting open sets.
\[ \mathbb{R} \setminus C_j \text{ open for each } j \]
Apply the finite intersection law
Why: A finite intersection of open sets is open, using the minimum-of-radii argument. So the complement of the union is open, which makes the union closed.
\[ \bigcap_{j=1}^{k} (\mathbb{R} \setminus C_j) \text{ open} \;\Rightarrow\; \bigcup_{j=1}^{k} C_j \text{ closed} \]
Verify on a concrete pair
Why: The union of the closed intervals from 0 to 1 and from 2 to 3 has complement equal to three open rays and gaps, all open, so the union is closed. The finiteness was essential: an infinite union of closed points, one at each reciprocal, is not closed since it misses zero.
\[ [0,1] \cup [2,3]:\ \mathbb{R} \setminus \big([0,1]\cup[2,3]\big) = (-\infty,0)\cup(1,2)\cup(3,\infty)\ \text{open} \]
Picture it
Animation
Shows: Each line of the worked example "a finite union of closed sets is closed", appearing one at a time.
The same working the example does, in the order a tutor would write it.
Takeaway: The union of the closed intervals from 0 to 1 and from 2 to 3 has complement equal to three open rays and gaps, all open, so the union is closed. The finiteness was essential: an infinite union of closed points, one at each reciprocal, is not closed since it misses zero.
Prediction
Predict first
What is this infinite intersection, and is it open?
Answer it in your own words, now, with nothing to choose from. The options are on the next slide — and picking the right one off a list is an easier skill than producing it.
Correct: It equals the half-open interval that includes 1 but not 0, which is not open.
Why: A point x is in every interval exactly when 0 is less than x and x is at most 1 plus 1 over n for all n. The infimum of those right ends is 1, and it is attained as a bound, so x can equal 1 but must exceed 0. The result is the interval from 0 to 1 including 1, which is not open because 1 has no room to its right.
Check
Consider the intersection over all positive integers of the open intervals from 0 to one plus one over n.
\[ \bigcap_{n \ge 1} \left(0,\ 1 + \tfrac{1}{n}\right) \]
Check your understanding
What is this infinite intersection, and is it open?
Answer: A
Why: A point x is in every interval exactly when 0 is less than x and x is at most 1 plus 1 over n for all n. The infimum of those right ends is 1, and it is attained as a bound, so x can equal 1 but must exceed 0. The result is the interval from 0 to 1 including 1, which is not open because 1 has no room to its right.
Concept
Between any two distinct reals sits a rational number. The rationals are dense in the line.
\[ \forall\, a < b \ \exists\, q \in \mathbb{Q}:\ a < q < b \]
Density says every neighborhood of every real, however tiny, contains a rational. This is the raw material for the next closure computation.
Ranking
Put in order
Put the moves of Worked example: the closure of the rationals is everything into the order they have to happen.
Why: These are the moves of the worked example in the order it makes them, and each one is set up by the one before it. To show a real is a limit point of the rationals, we must find a rational other than it in every neighborhood.
Worked example
Claim: the closure of the rationals is the entire real line. Every real is approached by rationals.
\[ \overline{\mathbb{Q}} = \mathbb{R} \]
Take an arbitrary real and any radius
Why: To show a real is a limit point of the rationals, we must find a rational other than it in every neighborhood.
\[ \text{fix } r \in \mathbb{R},\ \varepsilon > 0 \]
Density supplies a rational in the neighborhood
Why: The open interval from r minus epsilon to r plus epsilon has two distinct ends, so density gives a rational strictly inside it, and irrationals guarantee it differs from r when needed.
\[ \exists\, q \in \mathbb{Q}:\ r - \varepsilon < q < r + \varepsilon \]
Conclude every real is in the closure
Why: Every real is either a rational already in the set or a limit point of it, so the closure swallows the whole line.
\[ \mathbb{R} \subseteq \overline{\mathbb{Q}} \subseteq \mathbb{R} \]
Verify at a concrete irrational
Why: Approach the square root of two with the decimal rationals one point four, one point four one, one point four one four, and so on. Each is rational and they close in on the target, so the square root of two is a limit point of the rationals, as the claim requires.
\[ 1.4,\ 1.41,\ 1.414,\ 1.4142,\ \ldots \to \sqrt{2} \]
Picture it
Animation
Shows: Each line of the worked example "the closure of the rationals is everything", appearing one at a time.
The same working the example does, in the order a tutor would write it.
Takeaway: Approach the square root of two with the decimal rationals one point four, one point four one, one point four one four, and so on. Each is rational and they close in on the target, so the square root of two is a limit point of the rationals, as the claim requires.
Concept
A set is bounded when it fits inside some finite interval: there is a number past which no member lies on either side.
\[ S \text{ bounded} \iff \exists\, M > 0:\ S \subseteq [-M,\ M] \]
Boundedness controls size but says nothing about edges. We will see it is only half of what compactness needs.
Intuition
Compactness is the property that lets you replace an infinite amount of covering work with a finite amount. Whatever infinite family of open patches you throw at a compact set, finitely many already do the job.
Think of it as a guarantee against escaping to the edge or to infinity: the set is tightly held, no leak to slip through with only finitely many patches.
This is the same instinct as propositional compactness from the logic decks: an infinite set of constraints is satisfiable exactly when every finite piece is. Finiteness control, again.
Picture it
Figure (svg): A horizontal segment representing a set K on the number line, covered by five overlapping open intervals drawn as arcs above it. The arcs overlap so that together they span the whole segment, illustrating an open cover.
Discussion prompt
Read the picture before the words. What is this showing, and what is the one thing it is built to make obvious? Commit to an answer, then read on.
Hint: Name the parts, then say what changes between them — and if nothing changes, say what is being held still.
Answer:
An open cover of a set is a collection of open sets whose union contains the set. The patches together hide the whole thing.
Concept
An open cover of a set is a collection of open sets whose union contains the set. The patches together hide the whole thing.
\[ \{U_i\}_{i \in I} \text{ covers } K \iff K \subseteq \bigcup_{i \in I} U_i,\ \text{each } U_i \text{ open} \]
Figure (svg): A horizontal segment representing a set K on the number line, covered by five overlapping open intervals drawn as arcs above it. The arcs overlap so that together they span the whole segment, illustrating an open cover.
A subcover is a subcollection that still covers. The question compactness asks is whether you can always thin the cover down to a finite subcollection.
Concept
A set is compact when every open cover of it admits a finite subcover. No matter the cover, finitely many of its patches already suffice.
\[ K \text{ compact} \iff \forall \text{ open covers } \{U_i\}\ \exists\, i_1, \ldots, i_n:\ K \subseteq U_{i_1} \cup \cdots \cup U_{i_n} \]
compact set — A set such that every collection of open sets covering it has a finite subcollection that also covers it. This is the abstract, cover-based definition, valid in any topological space, not just the line.
The quantifier order is the crux: for every cover there exists a finite subcover. To disprove compactness you produce one cover with no finite subcover.
Hypothesis
Predict first
Worked example: the open unit interval is not compact is about to be worked. State your hypothesis first: which rule or definition decides this one, and what is the first move it forces? Then watch whether the example agrees with you.
Correct: Build the cover from the left-open pieces
Why: Each patch starts a little above zero and runs to one. As n grows the patch reaches closer to zero, so together they sweep the whole interval.
A hypothesis you wrote down is falsifiable; a vague sense of how it will go is not. If the example opens somewhere else, that gap is the thing worth chasing.
Worked example
Claim: the open interval from 0 to 1 is not compact. We exhibit one open cover that no finite piece can cover.
\[ \text{Show } (0,1) \text{ is not compact.} \]
Build the cover from the left-open pieces
Why: Each patch starts a little above zero and runs to one. As n grows the patch reaches closer to zero, so together they sweep the whole interval.
\[ U_n = \left(\tfrac{1}{n},\ 1\right), \quad n \ge 2 \]
Confirm the union covers the interval
Why: Any point x in the interval has a reciprocal below it once n is large enough, placing x inside that patch. So the patches cover everything strictly between 0 and 1.
\[ x \in (0,1) \Rightarrow \exists\, n:\ \tfrac{1}{n} < x \Rightarrow x \in U_n \]
Show no finite subfamily covers
Why: A finite subfamily has a largest index N, and its union is just the single widest patch, which starts at one over N. Points at or below one over N are left uncovered.
\[ U_{n_1} \cup \cdots \cup U_{n_k} = \left(\tfrac{1}{N},\ 1\right),\quad N = \max n_j \]
Verify a specific point is missed
Why: For any finite subfamily with largest index N, the point one over two N is positive, below one over N, hence in the interval but outside every chosen patch. So a finite subcover is impossible and the interval is not compact.
\[ \tfrac{1}{2N} \in (0,1),\quad \tfrac{1}{2N} < \tfrac{1}{N} \Rightarrow \tfrac{1}{2N} \notin \left(\tfrac{1}{N},1\right) \]
Picture it
Animation
Shows: Each line of the worked example "the open unit interval is not compact", appearing one at a time.
The same working the example does, in the order a tutor would write it.
Takeaway: For any finite subfamily with largest index N, the point one over two N is positive, below one over N, hence in the interval but outside every chosen patch. So a finite subcover is impossible and the interval is not compact.
Trap
The open interval from 0 to 1 fits neatly inside a finite window, so it is bounded, so it must be compact. Compactness is just smallness.
\[ (0,1) \subseteq [-1,1] \Rightarrow \text{bounded} \stackrel{?}{\Rightarrow} \text{compact} \]
Bounded is only half the story. The open interval is bounded yet not compact, because it is missing its endpoints and a cover can exploit the leak at zero.
Name the missing ingredient
Why: The cover by the intervals from one over n to one has no finite subcover precisely because the interval leaks out toward the absent endpoint zero. Boundedness plus closedness together is what compactness needs, not boundedness alone.
\[ (0,1)\ \text{bounded but not closed} \Rightarrow \text{not compact} \]
Step zero
Discussion prompt
Worked example: the whole line is not compact — before any calculation: what is the plan? Name the moves in order, in plain English, without doing the arithmetic.
Hint: It starts with: Cover the line by growing symmetric intervals
Answer:
Worked example
Claim: the real line is not compact. Here the leak is toward infinity rather than an endpoint.
\[ \text{Show } \mathbb{R} \text{ is not compact.} \]
Cover the line by growing symmetric intervals
Why: Each patch is the open interval from negative n to n. As n grows they swallow any point, so together they cover the whole line.
\[ V_n = (-n,\ n),\quad n \ge 1,\qquad \bigcup_n V_n = \mathbb{R} \]
A finite subfamily is bounded
Why: Finitely many of these patches have a largest index N, and their union is just the interval from negative N to N, a bounded set that cannot contain the unbounded line.
\[ V_{n_1} \cup \cdots \cup V_{n_k} = (-N,\ N),\quad N = \max n_j \]
Verify a point that escapes
Why: The number N itself lies outside the interval from negative N to N, so any finite subfamily misses it. No finite subcover exists, and the line is not compact. Unboundedness alone already defeats compactness.
\[ N \notin (-N,\ N) \Rightarrow \text{no finite subcover} \]
Picture it
Animation
Shows: Each line of the worked example "the whole line is not compact", appearing one at a time.
The same working the example does, in the order a tutor would write it.
Takeaway: The number N itself lies outside the interval from negative N to N, so any finite subfamily misses it. No finite subcover exists, and the line is not compact. Unboundedness alone already defeats compactness.
Concept
For subsets of the real line, the abstract cover condition matches a concrete geometric one exactly.
Heine-Borel: a subset of the line is compact if and only if it is closed and bounded, both together.
\[ K \subseteq \mathbb{R} \text{ compact} \iff K \text{ closed and bounded} \]
This is a theorem, not a definition. It turns a hard-to-check cover condition into a two-part checklist you can apply by eye.
Intuition
Closed and bounded forces compact only because the real line has no gaps. Completeness is the hidden engine.
Run the same test on the rationals: the rational points of the closed interval from 0 to the square root of two form a closed and bounded set there, yet a cover sneaking up on the missing square root of two has no finite subcover. The gap breaks it.
On the complete line there is no gap to leak through, so closed plus bounded really does trap the set. Order completeness becomes topological compactness.
Missing information
Discussion prompt
Use the closed-and-bounded checklist to sort four sets into compact or not compact.
What do you need to know — or decide — before the first line can be written? List everything the problem has to hand you.
Hint: Anything you would have to invent to get started is a thing the problem must supply.
Answer:
It is closed, since its complement is two open rays, and bounded, since it sits in a finite window. Both boxes ticked.
Worked example
Use the closed-and-bounded checklist to sort four sets into compact or not compact.
The closed unit interval is compact
Why: It is closed, since its complement is two open rays, and bounded, since it sits in a finite window. Both boxes ticked.
\[ [0,1]:\ \text{closed} + \text{bounded} \Rightarrow \text{compact} \]
The half-line from 0 upward is not compact
Why: It is closed but unbounded, so it fails the bounded half of the checklist.
\[ [0,\infty):\ \text{closed but unbounded} \Rightarrow \text{not compact} \]
The reciprocals alone are not compact
Why: The set of reciprocals is bounded but not closed, since it misses the limit point zero. Failing closed, it fails compact.
\[ \left\{\tfrac1n\right\}:\ \text{bounded but not closed} \Rightarrow \text{not compact} \]
Verify the repaired set is compact
Why: Adding zero to the reciprocals makes the set closed, since it now contains its only limit point, and it is still bounded. Closed plus bounded gives compact, confirming that patching the single missing limit point restores compactness.
\[ \left\{\tfrac1n\right\} \cup \{0\}:\ \text{closed} + \text{bounded} \Rightarrow \text{compact} \]
Picture it
Animation
Shows: Each line of the worked example "classify sets by Heine-Borel", appearing one at a time.
The same working the example does, in the order a tutor would write it.
Takeaway: Adding zero to the reciprocals makes the set closed, since it now contains its only limit point, and it is still bounded. Closed plus bounded gives compact, confirming that patching the single missing limit point restores compactness.
Check
Apply Heine-Borel to decide which of the following is compact.
Check your understanding
Which of these subsets of the real line is compact?
Answer: A
Why: The union of two closed intervals is closed, being a finite union of closed sets, and it is bounded, sitting inside the window from 0 to 3. Closed and bounded together give compact by Heine-Borel.
Concept
There is a second flavor of compactness phrased in sequences instead of covers.
A set is sequentially compact when every sequence drawn from it has a subsequence converging to a point of the set.
\[ K \text{ seq. compact} \iff \forall\, (x_n) \subseteq K\ \exists\ \text{subsequence } x_{n_k} \to L \in K \]
No sequence can wander off with all its energy escaping; something always clusters, and the cluster point stays inside.
Concept
For subsets of the real line, compact by covers and sequentially compact are the same property, both equal to closed and bounded.
\[ \text{compact} \iff \text{sequentially compact} \iff \text{closed and bounded} \]
The bridge is Bolzano-Weierstrass: every bounded sequence has a convergent subsequence, and closedness keeps the limit inside. The order completeness of the line powers both.
Estimation
Predict first
See sequential compactness fail on the open interval from 0 to 1, matching its failure by covers.
Commit before you compute: what does Worked example: a sequence leaking out of the open interval come out to? A rough magnitude and the right form is enough — the point is to have something concrete to be wrong about.
Correct: Verify against the checklist
Why: A prediction you can defend turns the computation into a check rather than a leap of faith — and an answer that contradicts it is caught on the spot. The open interval is bounded but not closed, so Heine-Borel already predicts failure, and the escaping sequence exhibits it directly.
Worked example
See sequential compactness fail on the open interval from 0 to 1, matching its failure by covers.
\[ \text{Show } (0,1) \text{ is not sequentially compact.} \]
Pick a sequence heading for the missing endpoint
Why: The reciprocals lie in the interval and march steadily toward zero, the endpoint the interval omits.
\[ x_n = \tfrac{1}{n} \in (0,1),\quad x_n \to 0 \]
Note every subsequence has the same limit
Why: A subsequence of a convergent sequence converges to the same limit, so every subsequence of the reciprocals also tends to zero.
\[ x_{n_k} \to 0 \text{ for every subsequence} \]
The limit is outside the set
Why: Zero is not in the open interval, so no subsequence converges to a point of the set. The defining condition of sequential compactness fails.
\[ 0 \notin (0,1) \]
Verify against the checklist
Why: The open interval is bounded but not closed, so Heine-Borel already predicts failure, and the escaping sequence exhibits it directly. Both compactness notions agree in refusing this set.
\[ (0,1)\ \text{not closed} \Rightarrow \text{not (sequentially) compact} \]
Picture it
Animation
Shows: Each line of the worked example "a sequence leaking out of the open interval", appearing one at a time.
The same working the example does, in the order a tutor would write it.
Takeaway: The open interval is bounded but not closed, so Heine-Borel already predicts failure, and the escaping sequence exhibits it directly. Both compactness notions agree in refusing this set.
Trap
Closed sounds like it should mean contained or limited, so a closed set ought to be finite, or at least bounded.
\[ \text{closed} \stackrel{?}{\Rightarrow} \text{finite or bounded} \]
Closed is about keeping limit points, and has nothing to do with size. Closed sets can be infinite and unbounded.
Exhibit closed sets that are large
Why: The whole line is closed and unbounded. The integers form a closed set, infinite and unbounded, with no limit points at all to miss. Closedness constrains edges, not extent, and only closed-and-bounded together yield compact.
\[ \mathbb{R},\ \mathbb{Z}:\ \text{closed, infinite, unbounded} \]
Two truths and a lie
Sort into buckets
Some of these hold up and some are the exact mistakes this lesson is built to prevent. Sort them.
Concept
A set is connected when it cannot be split into two nonempty pieces by open sets, each piece capturing part of it with no overlap.
\[ K \text{ disconnected} \iff \exists\ \text{open } U, V:\ K \subseteq U \cup V,\ U \cap K \neq \emptyset,\ V \cap K \neq \emptyset,\ U \cap V \cap K = \emptyset \]
On the real line the connected sets are exactly the intervals, in the widest sense: single points, rays, segments open or closed, and the whole line.
Intuition
An interval has no internal gap: if two of its points are in, everything between them is in as well. That is precisely what stops a clean two-way split.
A set with a gap can be cut at the gap by two opens, one on each side. So having a gap is the same as being disconnected, and having none is the same as being an interval.
Step zero
Discussion prompt
Worked example: a two-piece set is disconnected — before any calculation: what is the plan? Name the moves in order, in plain English, without doing the arithmetic.
Hint: It starts with: Cut in the gap between the pieces
Answer:
Worked example
Show the union of the closed intervals from 0 to 1 and from 2 to 3 is disconnected by exhibiting a separation.
\[ K = [0,1] \cup [2,3] \]
Cut in the gap between the pieces
Why: The number one and a half sits in the empty gap between the pieces. Splitting the line there gives two open half-lines, each catching one piece.
\[ U = (-\infty,\ 1.5),\qquad V = (1.5,\ \infty) \]
Check each piece lands in one part
Why: The first interval sits entirely below one and a half, the second entirely above, so each open set catches a nonempty part of the set.
\[ [0,1] \subseteq U,\qquad [2,3] \subseteq V \]
Confirm the separation conditions
Why: The two opens are disjoint, they cover the set, and each meets it, exactly the definition of a disconnection.
\[ U \cap V = \emptyset,\quad K \subseteq U \cup V,\quad U \cap K \neq \emptyset \neq V \cap K \]
Verify the cut point avoids the set
Why: One and a half belongs to neither interval, so the split runs cleanly through a genuine gap and no point of the set is stranded on the dividing line. The set is therefore disconnected, and being disconnected it is not an interval.
\[ 1.5 \notin K \Rightarrow \text{clean separation, } K \text{ disconnected} \]
Picture it
Animation
Shows: Each line of the worked example "a two-piece set is disconnected", appearing one at a time.
The same working the example does, in the order a tutor would write it.
Takeaway: The first interval sits entirely below one and a half, the second entirely above, so each open set catches a nonempty part of the set.
Pattern
1. Locate the leak
Why: Find where the set fails closed-and-bounded: a missing limit point at a finite edge, or an escape to infinity. That spot is where a cover will slip through.
2. Build a cover aimed at the leak
Why: Choose a family of open sets that approaches the leak but never quite plugs it, such as intervals creeping toward a missing endpoint or growing toward infinity.
3. Defeat every finite subfamily
Why: A finite subfamily has a widest member. Exhibit a point of the set beyond that member, proving no finite subcollection covers, hence the set is not compact.
\[ \text{no finite subcover} \Rightarrow \text{not compact} \]
Notation
Annotate
From Pattern: proving a set is not compact — read this one piece at a time. What is each part doing?
On: \( \text{no finite subcover} \Rightarrow \text{not compact} \)
Commit first
Predict first
Does this open cover of the closed unit interval have a finite subcover, and why?
Commit to an answer, then rate it — certain, fairly sure, or guessing — and write the rating down before you turn the page.
Correct: Yes: the interval is compact by Heine-Borel, so every open cover, including this one, has a finite subcover.
Why: The closed interval from 0 to 1 is closed and bounded, hence compact by Heine-Borel. Compactness means every open cover whatsoever has a finite subcover, so this particular cover does too; for instance a handful of centers spaced about one tenth apart already covers.
The rating matters as much as the answer: confident-and-wrong is the combination that survives revision, because nothing about it feels like it needs revisiting.
Check
Consider the closed interval from 0 to 1 covered by open intervals of radius one tenth centered at every point of the interval.
\[ \left\{\, \left(x - \tfrac{1}{10},\ x + \tfrac{1}{10}\right) : x \in [0,1] \,\right\} \]
Check your understanding
Does this open cover of the closed unit interval have a finite subcover, and why?
Answer: A
Why: The closed interval from 0 to 1 is closed and bounded, hence compact by Heine-Borel. Compactness means every open cover whatsoever has a finite subcover, so this particular cover does too; for instance a handful of centers spaced about one tenth apart already covers.
Concept
Every topological fact in this deck traces back to one order-theoretic fact: the completeness axiom, that every bounded set of reals has a least upper bound.
Completeness fills the gaps the rationals leave. With no gaps, closed-and-bounded sets cannot leak, so Heine-Borel holds; bounded sequences cannot wander off, so Bolzano-Weierstrass holds; and intervals cannot be split, so they are connected.
\[ \text{completeness} \Rightarrow \text{Heine-Borel},\ \text{Bolzano-Weierstrass},\ \text{connectedness} \]
Order became topology. The way you compare numbers dictated the shape of nearness on the line.
Concept
Compactness is the tool that trades an infinite job for a finite one, and that trade shows up far from analysis.
In the logic decks, propositional compactness said an infinite set of clauses is satisfiable exactly when every finite subset is. The pattern is identical: a global infinite property reduced to finitely many local checks.
Both are the same instinct dressed differently: finiteness control. It underlies termination arguments, cover-based resource proofs, and the leap from local rules to global guarantees across mathematics and computing.
Explain it
Discussion prompt
Explain The finiteness-control thread to computer science to a student a year behind you. No notation, no jargon they have not met — and it still has to be true.
Hint: If your explanation needs a symbol they have never seen, you are describing the notation rather than the idea.
Answer:
Compactness is the tool that trades an infinite job for a finite one, and that trade shows up far from analysis.
Prediction
Predict first
What is the closure of S, and is S itself compact?
Answer it in your own words, now, with nothing to choose from. The options are on the next slide — and picking the right one off a list is an easier skill than producing it.
Correct: The closure is S together with 0; S itself is not compact because it omits the limit point 0.
Why: The only limit point of the reciprocals is 0, since each reciprocal is isolated and the terms shrink below any radius. The closure adds exactly 0. Because S itself misses this limit point, S is not closed, so by Heine-Borel it is not compact even though it is bounded.
Check
Let S be the set of reciprocals of the positive integers. Decide its closure and whether it is compact.
\[ S = \left\{\, \tfrac{1}{n} : n \ge 1 \,\right\} \]
Check your understanding
What is the closure of S, and is S itself compact?
Answer: A
Why: The only limit point of the reciprocals is 0, since each reciprocal is isolated and the terms shrink below any radius. The closure adds exactly 0. Because S itself misses this limit point, S is not closed, so by Heine-Borel it is not compact even though it is bounded.
Intuition
Openness at every point has a clean structural payoff: an open set is nothing more than a union of open intervals.
Around each of its points sits a small open interval inside the set. Unioning all of those intervals rebuilds the set exactly, so every open set is a union of the basic open pieces.
\[ U \text{ open} \Rightarrow U = \bigcup_{x \in U} N_{\varepsilon_x}(x) \]
This is why open intervals are called a basis: they generate every open set by union, the way prime factors generate every integer by product.
Analogy
Discussion prompt
Explain Open sets are unions of intervals by analogy to something with no Foundations of Higher Mathematics in it at all — a queue, a recipe, a map, a bank balance, whatever fits. Then say where your analogy breaks.
Hint: An analogy that never breaks is not an analogy, it is the same idea wearing a hat. Find the seam — that is the part that is actually new.
Answer:
Openness at every point has a clean structural payoff: an open set is nothing more than a union of open intervals.
Concept
Two operations sit on either side of a set. The interior is the largest open set inside it; the closure is the smallest closed set containing it.
\[ \operatorname{int}(S) \subseteq S \subseteq \overline{S} \]
They are complementary: taking the interior of a set is the same as taking the closure of the complement and complementing again.
\[ \operatorname{int}(S) = \mathbb{R} \setminus \overline{\,\mathbb{R} \setminus S\,} \]
A set is open exactly when it equals its interior, and closed exactly when it equals its closure. Both extremes meet only for the clopen sets.
Counterexample
Discussion prompt
Two operations sit on either side of a set. The interior is the largest open set inside it; the closure is the smallest closed set containing it.
That is stated as though it always holds. Do one of two things: produce a case where it fails, or say precisely what rules such a case out. "It just does" is not on the menu.
Hint: Hunt at the extremes first — zero, one, negative, empty, equal. If every extreme survives, the reason they survive is the proof.
Answer:
They are complementary: taking the interior of a set is the same as taking the closure of the complement and complementing again.
Ranking
Put in order
Put the moves of Worked example: interior and closure of a half-open interval into the order they have to happen.
Why: These are the moves of the worked example in the order it makes them, and each one is set up by the one before it. The point zero has no room to its left inside the set, so it is not interior.
Worked example
Compute both the interior and the closure of the interval that includes 0 and excludes 1.
\[ S = [0,1) \]
Strip the non-interior endpoint for the interior
Why: The point zero has no room to its left inside the set, so it is not interior. Every point strictly between zero and one is interior, giving the open interval.
\[ \operatorname{int}(S) = (0,1) \]
Add the missing limit point for the closure
Why: The point one is a limit point of the set but not a member, so the closure adjoins it. Zero is already in, and no other points are approached.
\[ \overline{S} = [0,1] \]
Verify the two-sided sandwich
Why: The interior, the open interval, sits inside the set, which sits inside the closure, the closed interval, confirming the nesting. The set equals neither extreme, so it is neither open nor closed, matching the earlier verdict.
\[ (0,1) \subsetneq [0,1) \subsetneq [0,1] \]
Picture it
Animation
Shows: Each line of the worked example "interior and closure of a half-open interval", appearing one at a time.
The same working the example does, in the order a tutor would write it.
Takeaway: The interior, the open interval, sits inside the set, which sits inside the closure, the closed interval, confirming the nesting. The set equals neither extreme, so it is neither open nor closed, matching the earlier verdict.
Elimination
Eliminate the wrong options
Why must there be a real root between 1 and 2, and which topological property is responsible?
3 of these 4 are wrong. Strike them one at a time, and say what rules each one out before you strike the next. The survivor is the answer.
Survives elimination: A
Why: The Intermediate Value Theorem comes from connectedness: a continuous image of a connected set is connected, hence an interval, so it cannot jump from negative to positive without passing through 0. The sign change from f of 1 to f of 2 forces a root in between.
Check
A continuous function changes sign across the closed interval from 1 to 2. The interval is connected.
\[ f(x) = x^3 - x - 1,\quad f(1) = -1,\quad f(2) = 5 \]
Check your understanding
Why must there be a real root between 1 and 2, and which topological property is responsible?
Answer: A
Why: The Intermediate Value Theorem comes from connectedness: a continuous image of a connected set is connected, hence an interval, so it cannot jump from negative to positive without passing through 0. The sign change from f of 1 to f of 2 forces a root in between.
Connect it up
Draw it
One page, no notation unless you need it: draw how these connect — Pattern: classify any subset of the line · Pattern: proving a set is not compact · From order to topology · Open means wiggle room · The open interval. Put an arrow wherever one of them is what makes another possible, and label the arrow with why.
Recap
Topology of the line, built from the order in one move: distance from absolute value, then everything else.
Open means every point has wiggle room; closed means the complement is open, equivalently the set keeps all its limit points. The two are independent, so a set can be both (only the empty set and the whole line) or neither (a half-open interval).
| Property | Definition | Watch out |
|---|---|---|
| Open | every point has a neighborhood inside | arbitrary unions stay open, only finite intersections do |
| Closed | complement open / keeps limit points | not the opposite of open; can be infinite and unbounded |
| Compact | every open cover has a finite subcover | on the line: closed AND bounded, not bounded alone |
Compactness is finiteness control: Heine-Borel says compact equals closed and bounded on the line, powered by completeness, and it echoes propositional compactness from logic. To disprove compactness, exhibit one cover with no finite subcover.
Next: continuity, where preimages of open sets are open, carrying these topological properties across maps to give the Intermediate and Extreme Value Theorems.
Want this taught 1-on-1? Alexander tutors Foundations of Higher Mathematics — $55/session, free consultation.