Compose and UIs

Arrow provides several features which are very interesting when developing interactive applications, especially in combination with libraries with a similar functional flavor, such as Compose.

Example projects

Projects using Compose and Arrow can be found in the corresponding section.

Multiplatform-ready

All the libraries under the Arrow umbrella are Multiplatform-ready. This means you can use them in your Android applications using Jetpack Compose, and in Desktop, iOS, or Web alongside Compose Multiplatform.

Compose is functional

As opposed to other frameworks where stateful components are the norm, the architecture promoted by Compose brings many concepts traditionally associated with a more functional approach. For example, the UI is defined as a function taking as arguments the values from the current state. Updating the state is also explicitly marked, and often separated in a ViewModel.

As a consequence, Arrow and Compose make great dancing partners. Below we discuss a few feature which we think are an immediately gain for Android (and with Compose Multiplatform, other UI) developers. In the same vein, our design section readily applies to projects using Compose.

Simpler effectful code

Most applications don't live in a vacuum, they need to access other services or data sources. In those cases we write effectful code, where suspend and coroutines become relevant.

Arrow Fx introduces high-level concurrency as a way to simplify code where different actions must happen concurrently, ensuring that all rules of Structured Concurrency are followed, but without the hassle.

class UserSettingsModel: ViewModel() {
  private val _userData = mutableStateOf<UserData?>(null)
  val userData: State<UserData?> get() = _userData

  suspend fun loadUserData(userId: UserId) =
    parZip(
      { downloadAvatar(userId) },
      { UserRepository.getById(userId) }
    ) { avatarFile, user ->
      // this code is called once both finish
      _userData.value = UserData(
        id = userId,
        details = user,
        avatar = Avatar(avatarFile)
      )
    }
}

Anything outside your own application is the wilderness: connections are down, services are unavailable. Arrow's resilience module provides several ready-to-use patterns to better handle those situations, including retry policies and circuit breakers.

Built-in error types

One key part of every application is how the domain is modelled. Arrow emphasizes using immutable data. In particular, sealed hierarchies take the important role of describing the different states.

Although every application is unique, a common scenario in interactive applications involve having a "success state" and an "error state". For example, correctly loading the user data, or encountering a problem with connection or authentication. Instead of rolling your own types, Arrow (and our sibling library Quiver) provide out-of-the-box solutions:

• Either describes a model in which the application has either completely succeeded, or some amount of errors have occured. Validation is a prime example, since we usually require for all fields to be valid before moving forward with the data.
• Ior introduces a third option, namely succeeding but still with some problems along the way. This type is useful to model domains where we can work with some erroneous or missing information.
• Outcome models success, failure, and absence. The latter case is useful when the application may be in loading state: still no problems, but no data ready either.

Given the commonalities, Arrow provides a uniform API to work with values of those types.

Updating the model

One potential drawback of using immutable data to model your state is that updating it can become quite tiresome, because Kotlin provides no dedicated feature other than copy for this task.

class UserSettingsModel: ViewModel() {
  private val _userData = mutableStateOf<UserData?>(null)
  val userData: State<UserData?> get() = _userData

  fun updateName(
    newFirstName: String, newLastName: String
  ) {
    _userData.value = _userData.value.copy(
      details = _userData.value.details.copy(
        name = _userData.value.details.name.copy(
          firstName = newFirstName, lastName = newLastName
        )
      )
    )
  }
}

Arrow Optics addresses these drawbacks, providing tools for manipulating and transforming immutable data. The code above can be rewritten without boring repetition using the dedicated copy for MutableState.

class UserSettingsModel: ViewModel() {
  private val _userData = mutableStateOf<UserData?>(null)
  val userData: State<UserData?> get() = _userData

  fun updateName(
    newFirstName: String, newLastName: String
  ) {
    _userData.updateCopy {
      inside(UserData.details.name) {
        Name.firstName set newFirstName
        Name.lastName  set newLastName
      }
    }
  }
}