Style guide

Code styles shall be adhered to in the following order (whereas their precedence follows the order listed and subsequent guides can be used as fall back if the previous does not cover a specific topic):

1. This Guide
2. Android Kotlin Guides
3. JetBrain's Kotlin Coding Conventions

Naming

Variables

Variable names should be named to remove ambiguity. In general, including a variable's type in it's name is overly verbose, unless it provides necessary clarity. For example, a list of locations (List<Location>) would appropriately be named locations (by virtue of being plural, it implies that the variable is likely a Collection).

Data Streams

Data streams (e.g. Flow, Channel or LiveData) shall be named to describe the data flowing through them. For example, a data stream of singular color values (e.g. Flow<Color>) should be named color. For data streams of Collections (e.g. Flow<List<Color>>) they should take the plural form (e.g. colors).

Type	Naming	Example Type	Example Naming
Single value variable	Singular	Color	color
Collection	Plural	List<Color>	colors
Data flow of single values	Singular	Flow<Color>	color
Data flow of collections	Plural	Flow<List<Color>>	colors

Instances where naming would conflict (when you have a property to hold a single value of the same type as a property referencing a data flow, e.g. Color and Flow<Color>) the naming precedence shall follow the ordering of the table above, with the lower precedence being postfixed with it's type.

When you have naming conflicts with multiple data flow properties, their naming precedence is as follows:

• Flow
• Channel
• LiveData

interface Example {
    val color: Color // Single value property takes precedence over data flow properties.
    val colorFlow: Flow<Color> // Name postfixed with type to resolve naming conflict.
}

interface Example {
    val color: Flow<Color> // Flow takes precedence over LiveData.
    val colorLiveData: LiveData<Color> // Naming of LiveData property is postfixed to resolve naming conflict.
}

interface Example {
    val color: Color
    val colorFlow: Flow<Color>
    val colorLiveData: LiveData<Color>
}

Abbreviations

Abbreviations in variable names are strongly discouraged. Acronyms may be used if they are standard nomenclature (commonly used in place of their longhand counterparts).

val bookTitle = "Programming 101" // Okay

val bookTtl = "Programming 101" // WRONG!

val ipAddress = "127.0.0.1" // Okay

val ipAddr = "127.0.0.1" // WRONG!

Feature Flags

Boolean Feature Flags configured remotely should be named such that true coresponds to the feature being enabled and false is disabled.

val featureAbc = ExampleRemoteBoolean("feature_abc", true) // Okay, defaults to enabled

val featureAbc = ExampleRemoteBoolean("feature_abc", false) // Okay, defaults to disabled

val featureAbc = ExampleRemoteBoolean("feature_abc_disabled", true) // WRONG! true would result in feature being enabled

Android Resources

See Android Resources document for resource ID naming conventions.

Nullability

Operation	Usage ("When the value...")	Throws
!!	is "known" to be non-null.	KotlinNullPointerException
checkNotNull	is "expected" to be non-null.	IllegalStateException
requireNotNull	"must" not be null.	IllegalArgumentException

!!

The use of !! (double bang) operator is allowed but should be used sparingly. It should only be used when the value is known to never be null (but the compiler is unable to ascertain nullness), such as:

• Type provided by Java code does not have nullability markers
• Smart casting is not possible when accessing variable from another module

When the reason for the !! usage is not immediately obvious, then a comment should be added with rationale for the !! usage.

checkNotNull

checkNotNull should be used when a variable can be null but you are not expecting it to be. An example is having a nullable var that is inialized to null but is assigned a value at a later time. When accessing the var (and expecting the value to no longer be null) then checkNotNull should be used.

requireNotNull

When you are provided a value (for example, as an argument of a function) to which it's nullness is unknown (e.g. Java type that is missing nullability markers) but value must not be null, then requireNotNull should be used.

Coroutines

suspend vs CoroutineScope

Per Kotlin Coroutines in Practice by Roman Elizarov, the following conventions are recommended:

Paradigm	Usage
suspend	Does something long & waits for it to complete without blocking.
CoroutineScope function	Launches new Coroutines & quickly returns, does not wait for them.

// Does not return until it (and it's children) are done.
suspend fun downloadContent(location: Location): Content

// Launches Coroutines and returns immediately.
fun CoroutineScope.processReferences(...)

License

This work is licensed under a Creative Commons Attribution 4.0 International License.