Practical guide · verified against the real thing
What JSON actually is (and why every API speaks it)
In one line: The data format that quietly runs the modern web: objects, arrays, the strict little syntax rules, and the errors everyone makes.
JSON is how machines pass structured data around in plain text. An API answering "who is this user?", a configuration file, a browser saving your preferences — underneath, very often, is a snippet of JSON. It stands for JavaScript Object Notation, and its origin explains its shape: it is literally the way JavaScript writes its own objects, adopted by everything else because it is simple, human-readable and unambiguous.
The whole language in five lines
JSON has exactly a few building blocks: objects (curly braces holding "key": value pairs), arrays (square brackets holding an ordered list), strings (always double quotes), numbers, and the literals true, false, null. Nest them and you can describe almost anything: an object for a user containing an array of orders, each an object of its own. That is the entire language — which is precisely why it won. There is no versioning drama, no schema you must learn first; the format is defined by one short standards document, RFC 8259.
The strictness everyone trips on
JSON looks forgiving and is not. The rules that catch people daily: keys and strings use double quotes only — single quotes are invalid; no trailing commas — the last item in an object or array must not be followed by one; and comments do not exist in JSON, full stop (config formats that allow them, like YAML or JSON-with-comments flavours, are different formats). When a parser rejects your file, it is usually one of those three, and the error message points at the exact character — which is why the practical workflow is: paste, format, read the error position. The JSON formatter on this site does exactly that, in your browser.
Why it beat XML
The previous answer to this problem was XML — angle-bracketed, verbose, and heavy with ceremony. JSON describes the same data in a fraction of the characters, maps directly onto programming-language data structures, and needs no closing tags. The industry's migration was less a decision than a relief. XML survives where documents (not data) are the product; everywhere else, APIs speak JSON by default.
The two jobs to know
Parsing turns JSON text into live data structures your code can use; serialising turns your structures back into text to send or store. Every serious language has both built in (JSON.parse and JSON.stringify in JavaScript, the json module in Python). One security note that matters: parsing JSON from an untrusted source is reading data — the danger was never the format but what code does with it. And storing sensitive things? A format is not a lock: secrets in plain files are a mistake in any format.
Sources
Next