RightSignal
1 of 4

Your RightSignal

Concept of the day1 / 3

UTF-8

The concept

How one web page can mix English, accents, Chinese and emoji

Unicode gives every character a number, called a code point: 'A' is 65, 'é' is 233, '€' is 8364. UTF-8 is the rule for turning that number into bytes a file or network can store.

Small numbers get few bytes and big numbers get more, from 1 to 4 bytes. So plain English stays as compact as old ASCII, the 128-character code from the 1960s, while any other script still fits.

What it’s forDecides how many bytes UTF-8 uses for one character.

bytes(c) = 1 if c < 2⁷, 2 if c < 2¹¹, 3 if c < 2¹⁶, else 4

c
the character's Unicode code point, e.g. 233 for 'é'
2⁷
128: everything below this is plain ASCII
2¹¹
2,048: covers most accented Latin, Greek, Cyrillic, Arabic, Hebrew
2¹⁶
65,536: covers most Chinese, Japanese and Korean characters

What it givesA length from 1 to 4 bytes for each character.

  1. 'A' = 65 < 128 → 1 byte: 41
  2. 'é' = 233 < 2,048 → 2 bytes: C3 A9
  3. '€' = 8,364 < 65,536 → 3 bytes: E2 82 AC