JSON

Revision as of 12:55, 2 April 2015

JSON (JavaScript Object Notation) is widely used by web applications to communicate between a web page front end and a server back end. While it is named after JavaScript, there are libraries for many other programming and scripting languages to let them use this format as well. The geospatial format GeoJSON is based on JSON, as is the remote-procedure-call protocol JSON-RPC and the JSON-LD linked-data format. The IBM standard of JSONx is an XML implementation of JSON. I-JSON is a restricted profile of JSON for Internet use.

Format

JSON is a very simply-defined format for representing data in text-based form using a few simple types, including two "compound types" which can be nested:

• number: A decimal number, optionally including a decimal point or exponential notation. Leading zeroes are not allowed unless the only number before the decimal point is a single zero, or the number is equal to zero. (Other formats and programming languages may call these int, float, double, long, etc.)

• string: A string of (Unicode) characters surrounded by double quotes ("). The only characters not allowed directly (without escaping) are double quotes, backslashes, or control characters; a backslash can be used to precede a quote or backslash used as a character within the string, as well as \b for backspace, \f for formfeed, \n for newline, \r for carriage return, \t for tab, and \u followed by 4 hexadecimal digits for encoding any Unicode character in the 0000-FFFF range. (Other formats and programming languages may call these string or array of char).

• true, false, or nil are permitted as typeless values. (Other formats and programming languages may put these into boolean or pointer types.)

• object: An associative array consisting of names and values, where the name and value are separated by a colon (:), the name/value pairs are separated by commas, and the whole object is surrounded by curly braces {}. The name is a (double-quoted) string, and the value can be any of the data types in JSON including another object. There is no inherent ordering to the values of an object. (Other formats and programming languages may call these associative arrays, hashes, or dictionaries.)

• array: An ordered collection of values, which can each be any of the JSON data types, and are separated by commas and surrounded by square brackets []. (Other formats and programming languages may call these arrays or lists.)

JSONP

JSONP is a variant of JSON, "padded" with additional JavaScript code to permit the entire data set to be a valid piece of JS code able to be imported via a <script> statement for execution. Generally the extra code will be a function call encompassing the JSON data as its parameter, but sometimes it will be a variable assignment or other code. This gets around security restrictions in browser implementations of JavaScript which limit use of external data loaded from other sites (script statements can reference outside URLs), making it both very powerful and very dangerous if misused.

JSON Pointer

JSON Pointer (described in RFC 6901) is a syntax for referencing a specific value within a JSON document.

See also

• GeoJSON
• JSON API
• JSON-RPC
• JSONx

Specs and schemas

• Official website (includes spec and links to libraries in many languages)
• JSON Schema
• ECMA-404
• RFC 4627

Sample data

• Data from European Parliament in JSON form (Source: ParlTrack)

Utilities

• Linter/Validator
• Quick JSON pretty reformatter / JSON-tidy
• gojson: A simple command-line tool for manipulating JSON for use in developing Go code (generates Go structures from JSON)
• Go package to encode/decode JSON
• jq: command line JSON processor
• PapaParse: in-browser conversions between CSV and JSON
• Textract: extract text from various document formats including JSON

Other links

• Wikipedia: JSON
• Wikipedia: JSONP
• JSON Validation Roundup
• Why does Google prepend while(1); to their JSON responses?