Effective Dart (original) (raw)

Over the past several years, we've written a ton of Dart code and learned a lot about what works well and what doesn't. We're sharing this with you so you can write consistent, robust, fast code too. There are two overarching themes:

1. Be consistent. When it comes to things like formatting, and casing, arguments about which is better are subjective and impossible to resolve. What we do know is that being consistent is objectively helpful.
   If two pieces of code look different it should be because they are different in some meaningful way. When a bit of code stands out and catches your eye, it should do so for a useful reason.
2. Be brief. Dart was designed to be familiar, so it inherits many of the same statements and expressions as C, Java, JavaScript and other languages. But we created Dart because there is a lot of room to improve on what those languages offer. We added a bunch of features, from string interpolation to initializing formals, to help you express your intent more simply and easily.
   If there are multiple ways to say something, you should generally pick the most concise one. This is not to say you should code golf yourself into cramming a whole program into a single line. The goal is code that is_economical_, not dense.

We split the guidelines into a few separate pages for easy digestion:

• Style Guide – This defines the rules for laying out and organizing code, or at least the parts that dart format doesn't handle for you. The style guide also specifies how identifiers are formatted:camelCase, using_underscores, etc.
• Documentation Guide – This tells you everything you need to know about what goes inside comments. Both doc comments and regular, run-of-the-mill code comments.
• Usage Guide – This teaches you how to make the best use of language features to implement behavior. If it's in a statement or expression, it's covered here.
• Design Guide – This is the softest guide, but the one with the widest scope. It covers what we've learned about designing consistent, usable APIs for libraries. If it's in a type signature or declaration, this goes over it.

For links to all the guidelines, see thesummary.

How to read the guides

#

Each guide is broken into a few sections. Sections contain a list of guidelines. Each guideline starts with one of these words:

• DO guidelines describe practices that should always be followed. There will almost never be a valid reason to stray from them.
• DON'T guidelines are the converse: things that are almost never a good idea. Hopefully, we don't have as many of these as other languages do because we have less historical baggage.
• PREFER guidelines are practices that you should follow. However, there may be circumstances where it makes sense to do otherwise. Just make sure you understand the full implications of ignoring the guideline when you do.
• AVOID guidelines are the dual to "prefer": stuff you shouldn't do but where there may be good reasons to on rare occasions.
• CONSIDER guidelines are practices that you might or might not want to follow, depending on circumstances, precedents, and your own preference.

Some guidelines describe an exception where the rule does not apply. When listed, the exceptions may not be exhaustive—you might still need to use your judgement on other cases.

This sounds like the police are going to beat down your door if you don't have your laces tied correctly. Things aren't that bad. Most of the guidelines here are common sense and we're all reasonable people. The goal, as always, is nice, readable and maintainable code.

The Dart analyzer provides a linter to help you write good, consistent code that follows these and other guidelines. If one or more linter rules exist that can help you follow a guideline then the guideline links to those rules. The links use the following format:

Linter rule:unnecessary_getters_setters

To learn how to use the linter, see Enabling linter rules and the list of linter rules.

To keep the guidelines brief, we use a few shorthand terms to refer to different Dart constructs.

• A library member is a top-level field, getter, setter, or function. Basically, anything at the top level that isn't a type.
• A class member is a constructor, field, getter, setter, function, or operator declared inside a class. Class members can be instance or static, abstract or concrete.
• A member is either a library member or a class member.
• A variable, when used generally, refers to top-level variables, parameters, and local variables. It doesn't include static or instance fields.
• A type is any named type declaration: a class, typedef, or enum.
• A property is a top-level variable, getter (inside a class or at the top level, instance or static), setter (same), or field (instance or static). Roughly any "field-like" named construct.

Summary of all rules

#

Identifiers

• DO name types using UpperCamelCase.
• DO name extensions using UpperCamelCase.
• DO name packages, directories, and source files using lowercase_with_underscores.
• DO name import prefixes using lowercase_with_underscores.
• DO name other identifiers using lowerCamelCase.
• PREFER using lowerCamelCase for constant names.
• DO capitalize acronyms and abbreviations longer than two letters like words.
• PREFER using wildcards for unused callback parameters.
• DON'T use a leading underscore for identifiers that aren't private.
• DON'T use prefix letters.
• DON'T explicitly name libraries.

Ordering

• DO place dart: imports before other imports.
• DO place package: imports before relative imports.
• DO specify exports in a separate section after all imports.
• DO sort sections alphabetically.

Formatting

• DO format your code using dart format.
• CONSIDER changing your code to make it more formatter-friendly.
• PREFER lines 80 characters or fewer.
• DO use curly braces for all flow control statements.

Comments

• DO format comments like sentences.
• DON'T use block comments for documentation.

Doc comments

• DO use /// doc comments to document members and types.
• PREFER writing doc comments for public APIs.
• CONSIDER writing a library-level doc comment.
• CONSIDER writing doc comments for private APIs.
• DO start doc comments with a single-sentence summary.
• DO separate the first sentence of a doc comment into its own paragraph.
• AVOID redundancy with the surrounding context.
• PREFER starting comments of a function or method with third-person verbs if its main purpose is a side effect.
• PREFER starting a non-boolean variable or property comment with a noun phrase.
• PREFER starting a boolean variable or property comment with "Whether" followed by a noun or gerund phrase.
• PREFER a noun phrase or non-imperative verb phrase for a function or method if returning a value is its primary purpose.
• DON'T write documentation for both the getter and setter of a property.
• PREFER starting library or type comments with noun phrases.
• CONSIDER including code samples in doc comments.
• DO use square brackets in doc comments to refer to in-scope identifiers.
• DO use prose to explain parameters, return values, and exceptions.
• DO put doc comments before metadata annotations.

Markdown

• AVOID using markdown excessively.
• AVOID using HTML for formatting.
• PREFER backtick fences for code blocks.

Writing

• PREFER brevity.
• AVOID abbreviations and acronyms unless they are obvious.
• PREFER using "this" instead of "the" to refer to a member's instance.

Libraries

• DO use strings in part of directives.
• DON'T import libraries that are inside the src directory of another package.
• DON'T allow an import path to reach into or out of lib.
• PREFER relative import paths.

Null

• DON'T explicitly initialize variables to null.
• DON'T use an explicit default value of null.
• DON'T use true or false in equality operations.
• AVOID late variables if you need to check whether they are initialized.
• CONSIDER type promotion or null-check patterns for using nullable types.

Strings

• DO use adjacent strings to concatenate string literals.
• PREFER using interpolation to compose strings and values.
• AVOID using curly braces in interpolation when not needed.

Collections

• DO use collection literals when possible.
• DON'T use .length to see if a collection is empty.
• AVOID using Iterable.forEach() with a function literal.
• DON'T use List.from() unless you intend to change the type of the result.
• DO use whereType() to filter a collection by type.
• DON'T use cast() when a nearby operation will do.
• AVOID using cast().

Functions

• DO use a function declaration to bind a function to a name.
• DON'T create a lambda when a tear-off will do.

Variables

• DO follow a consistent rule for var and final on local variables.
• AVOID storing what you can calculate.

Members

• DON'T wrap a field in a getter and setter unnecessarily.
• PREFER using a final field to make a read-only property.
• CONSIDER using => for simple members.
• DON'T use this. except to redirect to a named constructor or to avoid shadowing.
• DO initialize fields at their declaration when possible.

Constructors

• DO use initializing formals when possible.
• DON'T use late when a constructor initializer list will do.
• DO use ; instead of {} for empty constructor bodies.
• DON'T use new.
• DON'T use const redundantly.

Error handling

• AVOID catches without on clauses.
• DON'T discard errors from catches without on clauses.
• DO throw objects that implement Error only for programmatic errors.
• DON'T explicitly catch Error or types that implement it.
• DO use rethrow to rethrow a caught exception.

Asynchrony

• PREFER async/await over using raw futures.
• DON'T use async when it has no useful effect.
• CONSIDER using higher-order methods to transform a stream.
• AVOID using Completer directly.
• DO test for Future when disambiguating a FutureOr whose type argument could be Object.

Names

• DO use terms consistently.
• AVOID abbreviations.
• PREFER putting the most descriptive noun last.
• CONSIDER making the code read like a sentence.
• PREFER a noun phrase for a non-boolean property or variable.
• PREFER a non-imperative verb phrase for a boolean property or variable.
• CONSIDER omitting the verb for a named boolean parameter.
• PREFER the "positive" name for a boolean property or variable.
• PREFER an imperative verb phrase for a function or method whose main purpose is a side effect.
• PREFER a noun phrase or non-imperative verb phrase for a function or method if returning a value is its primary purpose.
• CONSIDER an imperative verb phrase for a function or method if you want to draw attention to the work it performs.
• AVOID starting a method name with get.
• PREFER naming a method to___() if it copies the object's state to a new object.
• PREFER naming a method as___() if it returns a different representation backed by the original object.
• AVOID describing the parameters in the function's or method's name.
• DO follow existing mnemonic conventions when naming type parameters.

Libraries

• PREFER making declarations private.
• CONSIDER declaring multiple classes in the same library.

Classes and mixins

• AVOID defining a one-member abstract class when a simple function will do.
• AVOID defining a class that contains only static members.
• AVOID extending a class that isn't intended to be subclassed.
• DO use class modifiers to control if your class can be extended.
• AVOID implementing a class that isn't intended to be an interface.
• DO use class modifiers to control if your class can be an interface.
• PREFER defining a pure mixin or pure class to a mixin class.

Constructors

• CONSIDER making your constructor const if the class supports it.

Members

• PREFER making fields and top-level variables final.
• DO use getters for operations that conceptually access properties.
• DO use setters for operations that conceptually change properties.
• DON'T define a setter without a corresponding getter.
• AVOID using runtime type tests to fake overloading.
• AVOID public late final fields without initializers.
• AVOID returning nullable Future, Stream, and collection types.
• AVOID returning this from methods just to enable a fluent interface.

Types

• DO type annotate variables without initializers.
• DO type annotate fields and top-level variables if the type isn't obvious.
• DON'T redundantly type annotate initialized local variables.
• DO annotate return types on function declarations.
• DO annotate parameter types on function declarations.
• DON'T annotate inferred parameter types on function expressions.
• DON'T type annotate initializing formals.
• DO write type arguments on generic invocations that aren't inferred.
• DON'T write type arguments on generic invocations that are inferred.
• AVOID writing incomplete generic types.
• DO annotate with dynamic instead of letting inference fail.
• PREFER signatures in function type annotations.
• DON'T specify a return type for a setter.
• DON'T use the legacy typedef syntax.
• PREFER inline function types over typedefs.
• PREFER using function type syntax for parameters.
• AVOID using dynamic unless you want to disable static checking.
• DO use Future as the return type of asynchronous members that do not produce values.
• AVOID using FutureOr as a return type.

Parameters

• AVOID positional boolean parameters.
• AVOID optional positional parameters if the user may want to omit earlier parameters.
• AVOID mandatory parameters that accept a special "no argument" value.
• DO use inclusive start and exclusive end parameters to accept a range.

Equality

• DO override hashCode if you override ==.
• DO make your == operator obey the mathematical rules of equality.
• AVOID defining custom equality for mutable classes.
• DON'T make the parameter to == nullable.

Next Style

Was this page's content helpful?

Unless stated otherwise, the documentation on this site reflects Dart 3.10.3. Page last updated on 2025-9-4. View source or report an issue.