This is a living document and new ideas for improving the code around us are always welcome. Contribute: fork, clone, branch, commit, push, pull request.
- Rick Waldron @rwaldron, github
- Mathias Bynens @mathias, github
- Schalk Neethling @ossreleasefeed, github
- Kit Cambridge @kitcambridge, github
- Raynos github
- Matias Arriola @MatiasArriola, github
- John Fischer @jfroffice, github
- Idan Gazit @idangazit, github
- Leo Balter @leobalter, github
- Breno Oliveira @garu_rj, github
- Leo Beto Souza @leobetosouza, github
- Ryuichi Okumura @okuryu, github
- Pascal Precht @PascalPrecht, github
- EngForDev engfordev - Hwan Min Hong / MinTaek Kwon @leoinsight / Tw Shim @marocchino, github / Nassol Kim @nassol99, github / Juntai Park @rkJun, github / Minkyu Shim / Gangmin Won / Justin Yoo @justinchronicle / Daeyup Lee
- Marco Trulla @marcotrulla, github
- Alex Navasardyan @alexnavasardyan, github
All code in any code-base should look like a single person typed it, no matter how many people contributed.
The following list outlines the practices that I use in all code that I am the original author of; contributions to projects that I have created should follow these guidelines.
I do not intend to impose my style preferences on other people's code; if they have an existing common style - this should be respected.
"Part of being a good steward to a successful project is realizing that writing code for yourself is a Bad Idea™. If thousands of people are using your code, then write your code for maximum clarity, not your personal preference of how to get clever within the spec." - Idan Gazit
Leveraging Code Quality Tools by Anton Kovalyov
The following should be considered 1) incomplete, and 2) REQUIRED READING. I don't always agree with the style written by the authors below, but one thing is certain: They are consistent. Furthermore, these are authorities on the language.
- Baseline For Front End Developers
- Eloquent JavaScript
- JavaScript, JavaScript
- Adventures in JavaScript Development
- Perfection Kills
- Douglas Crockford's Wrrrld Wide Web
- JS Assessment
Projects should always attempt to include some generic means by which source can be linted, tested and compressed in preparation for production use. For this task, grunt by Ben Alman is second to none and has officially replaced the "kits/" directory of this repo.
Projects must include some form of unit, reference, implementation or functional testing. Use case demos DO NOT QUALIFY as "tests". The following is a list of test frameworks, none of which are endorsed more than the other.
- Whitespace
- Beautiful Syntax
- Type Checking (Courtesy jQuery Core Style Guidelines)
- Conditional Evaluation
- Practical Style
- Naming
- Misc
- Native & Host Objects
- Comments
- One Language Code
The following sections outline a reasonable style guide for modern JavaScript development and are not meant to be prescriptive. The most important take-away is the law of code style consistency. Whatever you chose as the style for your project should be considered law. Link to this document as a statement of your project's commitment to code style consistency, readability and maintainability.
- Never mix spaces and tabs.
- When beginning a project, before you write any code, choose between soft indents (spaces) or real tabs, consider this law.
- For readability, I always recommend setting your editor's indent size to two characters — this means two spaces or two spaces representing a real tab.
- If your editor supports it, always work with the "show invisibles" setting turned on. The benefits of this practice are:
- Enforced consistency
- Eliminating end of line whitespace
- Eliminating blank line whitespace
- Commits and diffs that are easier to read
-
A. Parens, Braces, Linebreaks
// if/else/for/while/try always have spaces, braces and span multiple lines // this encourages readability // 2.A.1.1 // Examples of really cramped syntax if(condition) doSomething(); while(condition) iterating++; for(var i=0;i<100;i++) someIterativeFn(); // 2.A.1.1 // Use whitespace to promote readability if ( condition ) { // statements } while ( condition ) { // statements } for ( var i = 0; i < 100; i++ ) { // statements } // Even better: var i, length = 100; for ( i = 0; i < length; i++ ) { // statements } // Or... var i = 0, length = 100; for ( ; i < length; i++ ) { // statements } var prop; for ( prop in object ) { // statements } if ( true ) { // statements } else { // statements }
B. Assignments, Declarations, Functions ( Named, Expression, Constructor )
// 2.B.1.1 // Variables var foo = "bar", num = 1, undef; // Literal notations: var array = [], object = {}; // 2.B.1.2 // Using only one `var` per scope (function) promotes readability // and keeps your declaration list free of clutter (also saves a few keystrokes) // Bad var foo = ""; var bar = ""; var qux; // Good var foo = "", bar = "", quux; // or.. var // Comment on these foo = "", bar = "", quux; // 2.B.1.3 // var statements should always be in the beginning of their respective scope (function). // Same goes for const and let from ECMAScript 6. // Bad function foo() { // some statements here var bar = "", qux; } // Good function foo() { var bar = "", qux; // all statements after the variables declarations. }
// 2.B.2.1 // Named Function Declaration function foo( arg1, argN ) { } // Usage foo( arg1, argN ); // 2.B.2.2 // Named Function Declaration function square( number ) { return number * number; } // Usage square( 10 ); // Really contrived continuation passing style function square( number, callback ) { callback( number * number ); } square( 10, function( square ) { // callback statements }); // 2.B.2.3 // Function Expression var square = function( number ) { // Return something valuable and relevant return number * number; }; // Function Expression with Identifier // This preferred form has the added value of being // able to call itself and have an identity in stack traces: var factorial = function factorial( number ) { if ( number < 2 ) { return 1; } return number * factorial( number-1 ); }; // 2.B.2.4 // Constructor Declaration function FooBar( options ) { this.options = options; } // Usage var fooBar = new FooBar({ a: "alpha" }); fooBar.options; // { a: "alpha" }
C. Exceptions, Slight Deviations
// 2.C.1.1 // Functions with callbacks foo(function() { // Note there is no extra space between the first paren // of the executing function call and the word "function" }); // Function accepting an array, no space foo([ "alpha", "beta" ]); // 2.C.1.2 // Function accepting an object, no space foo({ a: "alpha", b: "beta" }); // Single argument string literal, no space foo("bar"); // Inner grouping parens, no space if ( !("foo" in obj) ) { }
D. Consistency Always Wins
In sections 2.A-2.C, the whitespace rules are set forth as a recommendation with a simpler, higher purpose: consistency. It's important to note that formatting preferences, such as "inner whitespace" should be considered optional, but only one style should exist across the entire source of your project.
// 2.D.1.1 if (condition) { // statements } while (condition) { // statements } for (var i = 0; i < 100; i++) { // statements } if (true) { // statements } else { // statements }
E. Quotes
Whether you prefer single or double shouldn't matter, there is no difference in how JavaScript parses them. What ABSOLUTELY MUST be enforced is consistency. Never mix quotes in the same project. Pick one style and stick with it.
F. End of Lines and Empty Lines
Whitespace can ruin diffs and make changesets impossible to read. Consider incorporating a pre-commit hook that removes end-of-line whitespace and blanks spaces on empty lines automatically.
-
Type Checking (Courtesy jQuery Core Style Guidelines)
A. Actual Types
String:
typeof variable === "string"Number:
typeof variable === "number"Boolean:
typeof variable === "boolean"Object:
typeof variable === "object"Array:
Array.isArray( arrayLikeObject ) (wherever possible)Node:
elem.nodeType === 1null:
variable === nullnull or undefined:
variable == nullundefined:
Global Variables:
typeof variable === "undefined"Local Variables:
variable === undefinedProperties:
object.prop === undefined object.hasOwnProperty( prop ) "prop" in objectB. Coerced Types
Consider the implications of the following...
Given this HTML:
<input type="text" id="foo-input" value="1">
// 3.B.1.1 // `foo` has been declared with the value `0` and its type is `number` var foo = 0; // typeof foo; // "number" ... // Somewhere later in your code, you need to update `foo` // with a new value derived from an input element foo = document.getElementById("foo-input").value; // If you were to test `typeof foo` now, the result would be `string` // This means that if you had logic that tested `foo` like: if ( foo === 1 ) { importantTask(); } // `importantTask()` would never be evaluated, even though `foo` has a value of "1" // 3.B.1.2 // You can preempt issues by using smart coercion with unary + or - operators: foo = +document.getElementById("foo-input").value; // ^ unary + operator will convert its right side operand to a number // typeof foo; // "number" if ( foo === 1 ) { importantTask(); } // `importantTask()` will be called
Here are some common cases along with coercions:
// 3.B.2.1 var number = 1, string = "1", bool = false; number; // 1 number + ""; // "1" string; // "1" +string; // 1 +string++; // 1 string; // 2 bool; // false +bool; // 0 bool + ""; // "false"
// 3.B.2.2 var number = 1, string = "1", bool = true; string === number; // false string === number + ""; // true +string === number; // true bool === number; // false +bool === number; // true bool === string; // false bool === !!string; // true
// 3.B.2.3 var array = [ "a", "b", "c" ]; !!~array.indexOf("a"); // true !!~array.indexOf("b"); // true !!~array.indexOf("c"); // true !!~array.indexOf("d"); // false // Note that the above should be considered "unnecessarily clever" // Prefer the obvious approach of comparing the returned value of // indexOf, like: if ( array.indexOf( "a" ) >= 0 ) { // ... }
// 3.B.2.3 var num = 2.5; parseInt( num, 10 ); // is the same as... ~~num; num >> 0; num >>> 0; // All result in 2 // Keep in mind however, that negative numbers will be treated differently... var neg = -2.5; parseInt( neg, 10 ); // is the same as... ~~neg; neg >> 0; // All result in -2 // However... neg >>> 0; // Will result in 4294967294
-
// 4.1.1 // When only evaluating that an array has length, // instead of this: if ( array.length > 0 ) ... // ...evaluate truthiness, like this: if ( array.length ) ... // 4.1.2 // When only evaluating that an array is empty, // instead of this: if ( array.length === 0 ) ... // ...evaluate truthiness, like this: if ( !array.length ) ... // 4.1.3 // When only evaluating that a string is not empty, // instead of this: if ( string !== "" ) ... // ...evaluate truthiness, like this: if ( string ) ... // 4.1.4 // When only evaluating that a string _is_ empty, // instead of this: if ( string === "" ) ... // ...evaluate falsy-ness, like this: if ( !string ) ... // 4.1.5 // When only evaluating that a reference is true, // instead of this: if ( foo === true ) ... // ...evaluate like you mean it, take advantage of built in capabilities: if ( foo ) ... // 4.1.6 // When evaluating that a reference is false, // instead of this: if ( foo === false ) ... // ...use negation to coerce a true evaluation if ( !foo ) ... // ...Be careful, this will also match: 0, "", null, undefined, NaN // If you _MUST_ test for a boolean false, then use if ( foo === false ) ... // 4.1.7 // When only evaluating a ref that might be null or undefined, but NOT false, "" or 0, // instead of this: if ( foo === null || foo === undefined ) ... // ...take advantage of == type coercion, like this: if ( foo == null ) ... // Remember, using == will match a `null` to BOTH `null` and `undefined` // but not `false`, "" or 0 null == undefined
ALWAYS evaluate for the best, most accurate result - the above is a guideline, not a dogma.
// 4.2.1 // Type coercion and evaluation notes // Prefer `===` over `==` (unless the case requires loose type evaluation) // === does not coerce type, which means that: "1" === 1; // false // == does coerce type, which means that: "1" == 1; // true // 4.2.2 // Booleans, Truthies & Falsies // Booleans: true, false // Truthy: "foo", 1 // Falsy: "", 0, null, undefined, NaN, void 0
-
// 5.1.1 // A Practical Module (function( global ) { var Module = (function() { var data = "secret"; return { // This is some boolean property bool: true, // Some string value string: "a string", // An array property array: [ 1, 2, 3, 4 ], // An object property object: { lang: "en-Us" }, getData: function() { // get the current value of `data` return data; }, setData: function( value ) { // set the value of `data` and return it return ( data = value ); } }; })(); // Other things might happen here // expose our module to the global object global.Module = Module; })( this );
// 5.2.1 // A Practical Constructor (function( global ) { function Ctor( foo ) { this.foo = foo; return this; } Ctor.prototype.getFoo = function() { return this.foo; }; Ctor.prototype.setFoo = function( val ) { return ( this.foo = val ); }; // To call constructor's without `new`, you might do this: var ctor = function( foo ) { return new Ctor( foo ); }; // expose our constructor to the global object global.ctor = ctor; })( this );
-
A. You are not a human code compiler/compressor, so don't try to be one.
The following code is an example of egregious naming:
// 6.A.1.1 // Example of code with poor names function q(s) { return document.querySelectorAll(s); } var i,a=[],els=q("#foo"); for(i=0;i<els.length;i++){a.push(els[i]);}
Without a doubt, you've written code like this - hopefully that ends today.
Here's the same piece of logic, but with kinder, more thoughtful naming (and a readable structure):
// 6.A.2.1 // Example of code with improved names function query( selector ) { return document.querySelectorAll( selector ); } var idx = 0, elements = [], matches = query("#foo"), length = matches.length; for ( ; idx < length; idx++ ) { elements.push( matches[ idx ] ); }
A few additional naming pointers:
// 6.A.3.1 // Naming strings `dog` is a string // 6.A.3.2 // Naming arrays `dogs` is an array of `dog` strings // 6.A.3.3 // Naming functions, objects, instances, etc camelCase; function and var declarations // 6.A.3.4 // Naming constructors, prototypes, etc. PascalCase; constructor function // 6.A.3.5 // Naming regular expressions rDesc = //; // 6.A.3.6 // From the Google Closure Library Style Guide functionNamesLikeThis; variableNamesLikeThis; ConstructorNamesLikeThis; EnumNamesLikeThis; methodNamesLikeThis; SYMBOLIC_CONSTANTS_LIKE_THIS;
B. Faces of
thisBeyond the generally well known use cases of
callandapply, always prefer.bind( this )or a functional equivalent, for creatingBoundFunctiondefinitions for later invocation. Only resort to aliasing when no preferable option is available.// 6.B.1 function Device( opts ) { this.value = null; // open an async stream, // this will be called continuously stream.read( opts.path, function( data ) { // Update this instance's current value // with the most recent value from the // data stream this.value = data; }.bind(this) ); // Throttle the frequency of events emitted from // this Device instance setInterval(function() { // Emit a throttled event this.emit("event"); }.bind(this), opts.freq || 100 ); } // Just pretend we've inherited EventEmitter ;)
When unavailable, functional equivalents to
.bindexist in many modern JavaScript libraries.// 6.B.2 // eg. lodash/underscore, _.bind() function Device( opts ) { this.value = null; stream.read( opts.path, _.bind(function( data ) { this.value = data; }, this) ); setInterval(_.bind(function() { this.emit("event"); }, this), opts.freq || 100 ); } // eg. jQuery.proxy function Device( opts ) { this.value = null; stream.read( opts.path, jQuery.proxy(function( data ) { this.value = data; }, this) ); setInterval( jQuery.proxy(function() { this.emit("event"); }, this), opts.freq || 100 ); } // eg. dojo.hitch function Device( opts ) { this.value = null; stream.read( opts.path, dojo.hitch( this, function( data ) { this.value = data; }) ); setInterval( dojo.hitch( this, function() { this.emit("event"); }), opts.freq || 100 ); }
As a last resort, create an alias to
thisusingselfas an Identifier. This is extremely bug prone and should be avoided whenever possible.// 6.B.3 function Device( opts ) { var self = this; this.value = null; stream.read( opts.path, function( data ) { self.value = data; }); setInterval(function() { self.emit("event"); }, opts.freq || 100 ); }
C. Use
thisArgSeveral prototype methods of ES 5.1 built-ins come with a special
thisArgsignature, which should be used whenever possible// 6.C.1 var obj; obj = { f: "foo", b: "bar", q: "qux" }; Object.keys( obj ).forEach(function( key ) { // |this| now refers to `obj` console.log( this[ key ] ); }, obj ); // <-- the last arg is `thisArg` // Prints... // "foo" // "bar" // "qux"
thisArgcan be used withArray.prototype.every,Array.prototype.forEach,Array.prototype.some,Array.prototype.map,Array.prototype.filter -
This section will serve to illustrate ideas and concepts that should not be considered dogma, but instead exists to encourage questioning practices in an attempt to find better ways to do common JavaScript programming tasks.
A. Using
switchshould be avoided, modern method tracing will blacklist functions with switch statementsThere seems to be drastic improvements to the execution of
switchstatements in latest releases of Firefox and Chrome. http://jsperf.com/switch-vs-object-literal-vs-moduleNotable improvements can be witnesses here as well: rwaldron#13
// 7.A.1.1 // An example switch statement switch( foo ) { case "alpha": alpha(); break; case "beta": beta(); break; default: // something to default to break; } // 7.A.1.2 // A better approach would be to use an object literal or even a module: var switchObj = { alpha: function() { // statements // a return }, beta: function() { // statements // a return }, _default: function() { // statements // a return } }; var switchModule = (function () { return { alpha: function() { // statements // a return }, beta: function() { // statements // a return }, _default: function() { // statements // a return } }; })(); // 7.A.1.3 // If `foo` is a property of `switchObj` or `switchModule`, execute as a method... ( Object.hasOwnProperty.call( switchObj, foo ) && switchObj[ foo ] || switchObj._default )( args ); ( Object.hasOwnProperty.call( switchObj, foo ) && switchModule[ foo ] || switchModule._default )( args ); // If you know and trust the value of `foo`, you could even omit the OR check // leaving only the execution: switchObj[ foo ]( args ); switchModule[ foo ]( args ); // This pattern also promotes code reusability.
B. Early returns promote code readability with negligible performance difference
// 7.B.1.1 // Bad: function returnLate( foo ) { var ret; if ( foo ) { ret = "foo"; } else { ret = "quux"; } return ret; } // Good: function returnEarly( foo ) { if ( foo ) { return "foo"; } return "quux"; }
-
The basic principle here is:
To reinforce this concept, please watch the following presentation:
<iframe src="http://blip.tv/play/g_Mngr6LegI.html" width="480" height="346" frameborder="0" allowfullscreen></iframe>http://blip.tv/jsconf/jsconf2011-andrew-dupont-everything-is-permitted-extending-built-ins-5211542
- Single line above the code that is subject
- Multiline is good
- End of line comments are prohibited!
- JSDoc style is good, but requires a significant time investment
-
Programs should be written in one language, whatever that language may be, as dictated by the maintainer or maintainers.
Any project that cites this document as its base style guide will not accept comma first code formatting, unless explicitly specified otherwise by that project's author.