Uniform continuity — difference from pointwise

WHAT are we comparing?

HOW the gap appears — derive it from the slope

Take $f(x)=x^2$ on $\mathbb{R}$ . Let's derive why $\delta$ must shrink as $x$ grows.

For points $x,y$ : $|f(x)-f(y)| = |x^2-y^2| = |x+y|\,|x-y|.$

Why this step? Factoring exposes the amplifying factor $|x+y|$ : the same input gap $|x-y|$ produces a bigger output gap where $|x+y|$ is large.

To force $|x^2-y^2|<\varepsilon$ we'd need $|x-y| < \dfrac{\varepsilon}{|x+y|}$ . As $|x+y|\to\infty$ this required gap $\to 0$ . So no single $\delta>0$ can work for all $x,y$ :

The saving theorem (WHY compactness rescues us)

Sketch of WHY (first principles): Continuity gives a local $\delta_{x}$ at each point. The balls $(x-\delta_x/2,\,x+\delta_x/2)$ cover $[a,b]$ . Compactness extracts a finite subcover. Take the minimum of finitely many $\delta$ 's (a Lebesgue-number argument) — a positive number $\delta$ that works everywhere. On $\mathbb{R}$ (not compact) the inf of infinitely many shrinking $\delta$ 's can be $0$ , which is why $x^2$ fails.

Figure — Uniform continuity — difference from pointwise

A clean sufficient test (Lipschitz)

Formula Lipschitz

$\Rightarrow$ uniformly continuous If there is $L\ge 0$ with $|f(x)-f(y)|\le L|x-y|$ for all $x,y\in A$ , then $f$ is uniformly continuous: just take $\delta=\varepsilon/L$ (for $L>0$ ). Derivation: $|x-y|<\delta=\varepsilon/L \Rightarrow |f(x)-f(y)|\le L|x-y| < L\cdot\varepsilon/L=\varepsilon.$ $\delta$ has no $x$ in it — that's uniformity.

A bounded derivative $|f'|\le L$ gives Lipschitz by the Mean Value Theorem: $|f(x)-f(y)|=|f'(c)||x-y|\le L|x-y|$ .

Common mistakes (steel-manned)

Flashcards

Pointwise vs uniform — what is the single structural difference?

Quantifier order: pointwise allows

\delta=\delta(\varepsilon,x_0)

(chosen after the point); uniform requires

\delta=\delta(\varepsilon)

(one

\delta

for all points).

Write the uniform continuity definition.

\forall\varepsilon>0\,\exists\delta>0\,\forall x,y\in A:\;|x-y|<\delta\Rightarrow|f(x)-f(y)|<\varepsilon.

Why is

x^2

not uniformly continuous on

\mathbb{R}

|x^2-y^2|=|x+y||x-y|

; the amplifier

|x+y|

is unbounded, so no fixed

\delta

keeps the output gap

<\varepsilon

everywhere.

State Heine–Cantor.

Continuous on a compact set (closed & bounded

[a,b]

)

\Rightarrow

uniformly continuous there.

Why does compactness give uniform continuity?

Local

\delta

-balls cover; finite subcover; take the minimum

\delta>0

(Lebesgue number). On non-compact sets the inf can be

0

Lipschitz

\Rightarrow

uniform continuity: what

\delta

works?

\delta=\varepsilon/L

, since

|f(x)-f(y)|\le L|x-y|<\varepsilon

Give a uniformly continuous function with unbounded derivative.

\sqrt x

[0,\infty)

; use

|\sqrt x-\sqrt y|\le\sqrt{|x-y|}

, take

\delta=\varepsilon^2

Why does continuity fail to give uniform continuity on

(0,1)

(0,1)

isn't closed/compact; e.g.

1/x

blows up near

0

, forcing

\delta\to0

Recall Feynman: explain to a 12-year-old

Imagine drawing a graph and a rule: "if two dots are closer than $\delta$ sideways, they're closer than $\varepsilon$ up–down." Ordinary continuity lets you use a different sideways-rule in each region of the page — tiny steps where the curve is steep, big steps where it's flat. Uniform continuity says: I can pick one sideways-rule that works on the whole page at once. For a steep-forever curve like $x^2$ , the curve gets so steep far out that no single rule survives — so it's continuous but not uniformly so. On a closed, finite stretch the curve can't get infinitely steep, so one rule always exists.

Connections

Continuity (pointwise) — the weaker condition this strengthens.
Heine–Cantor Theorem — compactness upgrades continuity to uniform.
Compactness / Lebesgue number lemma — engine behind the upgrade.
Lipschitz continuity — a strong sufficient condition.
Mean Value Theorem — links bounded derivative to Lipschitz.
Cauchy sequences — uniform continuity preserves Cauchyness; pointwise need not.

Concept Map

Hinglish (regional understanding)

Intuition Hinglish mein samjho

Dekho, pointwise continuity aur uniform continuity ka asli farq sirf ek cheez hai — quantifier ka order, yaani $\delta$ kis cheez pe depend karta hai. Pointwise me har point $x_0$ ke liye apna alag $\delta$ choose kar sakte ho: jahan curve steep hai wahan chhota $\delta$ , jahan flat hai wahan bada $\delta$ . Uniform me condition strong hai — ek hi $\delta$ poore domain ke liye kaam karna chahiye, before kisi point ko dekhe.

Iska best example $x^2$ hai. $|x^2-y^2|=|x+y|\,|x-y|$ . Yahan $|x+y|$ ek "amplifier" hai. $\mathbb{R}$ pe jaise jaise $x$ bada hota hai, same input gap se output gap bohot bada ho jata hai. Isliye koi fixed $\delta$ kaam nahi karta — woh continuous hai par uniformly continuous nahi. Lekin agar tum $[0,5]$ jaisa closed-bounded (compact) interval lo, to $|x+y|\le 10$ bound ho jata hai, aur $\delta=\varepsilon/10$ sabke liye chal jata hai. Yahi Heine–Cantor theorem kehta hai: compact set pe continuity automatically uniform ban jati hai.

Do common galtiyan yaad rakho. Pehli: "continuous everywhere means uniformly continuous" — galat, kyunki unbounded domain pe $\delta$ shrink karke $0$ ho sakta hai. Doosri: " $(0,1)$ bounded hai to bas ho gaya" — nahi, $(0,1)$ closed nahi hai, isliye $1/x$ wahan continuous hai par uniformly continuous nahi (zero ke paas blow up). Aur ek surprise: $\sqrt{x}$ ka derivative $0$ ke paas infinite hai phir bhi woh uniformly continuous hai ( $\delta=\varepsilon^2$ lo) — matlab Lipschitz hona sufficient hai, necessary nahi. Mantra: "Uniform = ONE delta for everyone."

Uniform continuity — difference from pointwise

WHAT are we comparing?

HOW the gap appears — derive it from the slope

The saving theorem (WHY compactness rescues us)

A clean sufficient test (Lipschitz)

Common mistakes (steel-manned)

Flashcards

Connections

Concept Map

Hinglish (regional understanding)

Go deeper — visual, from zero

Test yourself — Advanced Topics (Elite Level)

Connections