Why Does JSON Disallow Trailing Commas? (RFC 8259 Deep Dive)

The Grammar Behind the Restriction

The official **IETF RFC 8259** (and preceding RFC 4627) specifies the formal Backus-Naur Form (BNF) grammar for JSON array production rules as follows:

array = begin-array [ value *( value-separator value ) ] end-array
value-separator = ws %x2C ws ; comma (",")

Notice the rule `*( value-separator value )`. This mathematical formulation explicitly dictates that every comma **must** be immediately followed by another valid `value`. If an ending bracket (`]` or `}`) is reached right after a comma, the grammar expects a value that does not exist, triggering an unrecoverable parser state.

The Sparse Array Ambiguity In JavaScript, an array initialized as `[1, 2, , 4]` creates a **sparse array** where index 2 is an empty elision slot (a hole, different from `undefined`).

If JSON allowed trailing commas without strict value anchoring, serializers in C, Go, or Java—which lack the concept of sparse memory arrays—would face cross-language compatibility ambiguities: - Should `["cat", "dog",]` allocate 2 or 3 elements? - Does a trailing comma signify an intentional empty element or an accidental typo?

To eliminate ambiguity across all hardware architectures, Crockford standardized commas strictly as separators.

Removing Trailing Commas with Regex When sanitizing untrusted human input on the client before calling `JSON.parse()`, you can remove trailing commas using this regex pattern:

function sanitizeJson(raw) {
  return raw.replace(/,\s*([}\]])/g, '$1');
}

You can also paste your raw text into Mickytool's **JSON Formatter & Validator** to automatically detect, flag, and remove malformed trailing characters in real time.