HTML entities reference

Three ways to write the same character

Every entity has up to three forms, and all of them produce an identical character in the rendered page.

Form	Example	When to prefer it
Named	—	HTML, where readability in the source matters
Decimal	—	XML, and anywhere the name might not be defined
Hexadecimal	—	When cross-checking against a Unicode chart, since U+2014 is the same number

HTML5 defines around 2,000 named entities. XML defines exactly five. That difference is why markup copied from an HTML context breaks an XML parser — see XML escape.

What actually has to be escaped

Far less than most people escape. In a UTF-8 HTML document:

• In text content: & and <. That is all.
• In an attribute value: those two plus whichever quote character delimits the attribute.
• Everywhere else: nothing. Type é, —, © and 你好 directly.

Escaping every accented character to a numeric reference is a habit left over from the era of ambiguous encodings. It triples the byte count, makes the source unreadable, and breaks copy-paste out of the source file. Declare <meta charset="utf-8"> and type the characters.

The invisible ones

Entity	What it does	Where it earns its place
 	A space that will not break	“10 kg”, “Figure 4”, preventing a lone short word at the end of a heading
 	A narrow space	Before French punctuation, and inside large numbers
‌	Prevents a ligature	Persian and Arabic typography; separating “fi” where a ligature is wrong
‍	Forces a join	Glues emoji sequences together — this is how the family emoji is built
­	A hyphenation point	Only renders as a hyphen if the line actually breaks there

Two warnings about  . It is not a normal space, so a string comparison or a regex looking for \s in some engines will not match it — this is a frequent source of “why does this lookup fail” bugs when content has been pasted from Word. And using a run of them for indentation is a layout hack; use CSS padding or margin.

Dashes and quotes

The four horizontal lines are genuinely different characters with different jobs, and using the right one is most of what separates typeset text from typed text.

• Hyphen - — compound words: well-known, twenty-one.
• En dash – — ranges and relationships: 1914–1918, the London–Paris route.
• Em dash — — a parenthetical break in a sentence.
• Minus − — arithmetic. Wider than a hyphen and aligned with the digits.

For quotes, the apostrophe in prose is ’ (’), not the vertical typewriter mark. Keep the vertical form for code, where it is a delimiter.

Why an entity sometimes shows as literal text

1. Double escaping. The ampersand itself was escaped upstream, so the browser receives   and correctly renders   as text.
2. A missing semicolon. HTML5 tolerates it for a handful of legacy entities and not for the rest.
3. The wrong context. Entities are processed in HTML text and attribute values. They are not processed inside a <textarea> value set from JavaScript, inside a JSON string, in a CSS content property (which uses \2014 instead), or in an XML document where the name is undeclared.
4. A framework escaped it for you. React, Vue and most template engines escape text by default, so writing an entity in a JSX string literal outputs it verbatim. Use the actual character, or a JavaScript escape such as \u2014.