Splitting APIs, Servers, and Implementations in Elixir

code better • have fun

tldr; I think the conventional way of structuring Elixir code could be improved by paying more attention to decoupling.

I just finished writing the first of my Coding Gnome courses. This one was an introduction to Elixir for experienced programmers.

I tried to concentrate on partitioning code in a reasonable manner. I didn’t use the traditional Elixir scheme, which comes from a mating of Ruby and Erlang project layouts. Instead, I tried to come at it with a fresh eye, asking myself how the various aspects of the code could best be decoupled.

Separating Execution Strategy from Logic

Elixir and Erlang have an interesting execution module. You program using processes and message passing, but they abstract this into the concept of servers. You call a function (typically GenServer.call(pid, args)) and this in turn send a message to the server identified by pid. Inside that server, you write callback functions that are invoked in response to these messages.

In real life, no one wants to use a server whose API involves such convolution. So the convention arose that you’d provide an API layer to your server, written in the same module. Here’s an example, stolen from the Elixir guide, and cut down somewhat:

defmodule KV.Registry do
  use GenServer

  ##############
  # Client API #
  ##############

  def start_link do
    GenServer.start_link(__MODULE__, :ok, [])
  end

  def lookup(server, name) do
    GenServer.call(server, {:lookup, name})
  end

  def create(server, name, value) do
    GenServer.call(server, {:store, name, value})
  end

  ####################
  # Server Callbacks #
  ####################
  
  def init(:ok) do
    {:ok, %{}}
  end

  def handle_call({:lookup, name}, _from, names) do
    {:reply, Map.fetch(names, name), names}
  end

  def handle_call({:store, name, value}, _from, names) do
    {:reply, Map.put(names, name, value)}
  end
end

Here the top half of the module is the public facing API, and the lower half is the code that runs in a separate process and that implements the functionality.

I’ve never been comfortable with this. It seems to bury the important part—the actual implementation—in amongst all kinds of GenServer housekeeping. It also makes the development of the code a lot trickier—you have to write the server at the same time that you write the implementation.

So in the course I recommended a different approach—one that I’ve been using personally for a while. In it, I write the application functionality in its own module, under lib/. This has no GenServer support—after all, it’s just the app logic.

defmodule Kv.Impl do       # in lib/kv/impl.ex
  def new() do
    %{}
  end

  def lookup(names, name) do
    Map.fetch(names, name)
  end

  def store(names, name, value) do
    Map.put(names, name, value)
  end
end

That makes it easier to see what’s going on. I can also write tests directly against this logic.

Now, here’s the weird part. You might look at this and say that here’s the module that other applications should call. But I think that’s not the case. Let’s instead declare the API in the top-level kv.ex file:

lib
├── kv
│   └── impl.ex
└── kv.ex               # <- the API belongs in here

Right now, this API just calls directly down to the implementation.

defmodule Kv do
  defdelegate new(),                     to: Kv.Impl
  defdelegate lookup(names, name),       to: Kv.Impl
  defdelegate store(names, name, value), to: Kv.Impl
end

Also pretty clean, right?

So now we have a working application (aka library), and people can start using it.

Bring On the Server

Circumstances change, and our library needs to become a full server.

We write the server code in lib/server.ex.

defmodule Kv.Server do
  use GenServer
  alias Kv.Impl

  def init(store) do
    { :ok, store }
  end

  def handle_call({:lookup, name}, _, store) do
    result = Impl.lookup(store, name)
    { :reply, result, store }
  end

  def handle_cast({:store, name, value}, store) do
    { :noreply, Impl.store(store, name, value) }
  end
end

This is just pure server code—no API, and no application logic.

Finally, we change the API in the top-level kv.ex file:

defmodule Kv do
  def new() do
    { :ok, names } = GenServer.start_link(Kv.Server, %{})
    names
  end
  
  def lookup(names, name) do
    GenServer.call(names, {:lookup, name})
  end
  
  def store(names, name, value) do
    GenServer.cast(names, { :store, name, value })
    names
  end
end

The only slightly strange thing is that store() returns the server pid. Doing so maintains the same API that we had previously, where the store() function returned the updated map. In both cases, the return value simply represents an (opaque) updated state.

As a result, we now have:

lib
├── kv
│   ├── impl.ex . . . . . . .   Application implementation
│   └── server.ex . . . . . .   GenServer implementation
└── kv.ex . . . . . . . . . .   API

This is the scheme I now use, and so far I much prefer it to the conventional one.

Bonus Section

Subservers

What if my application needs its own GenServers as part of its implementation? Well, just follow the same pattern, but one level down the directory tree:

lib
├── kv
│   ├── bucket
│   │   ├── impl.ex
│   │   └── server.ex
│   ├── bucket.ex
│   ├── impl.ex
│   └── server.ex
└── kv.ex

The rule here is that no one outside the application is allowed to call functions outside lib/kv.ex.

Applications are Components

Although the good folks who brought us Erlang were really, really smart people who produced designs for a future that did not yet exist, they really weren’t that good at naming things.

One of the most confusing names is application.

In “the real world” an application is something you deliver to the end user. A payroll system is an application. A word processor is an application.

But in Erlang, and hence Elixir, an application is a self-contained bundle of modules and resources with its own lifecycle. The Logger, for example, is an application. The Elixir compiler contains over a dozen.

Erlang applications are really just components.

But because the word application is so ingrained in us developers, it is hard to remember this. And so we have a tendency to throw all our code into a single project tree because, after all, it’s the application.

So I’m trying to retrain my brain by writing my code as series of separate applications, each as small as I can make it. And I’m not using umbrella projects for this, because I want to be able to share these components across different projects. Instead, I just use file dependencies.

So, next time you’re writing a killer Phoenix app, think about why you have Ecto in your web tier. Shouldn’t the business logic be out in its own application? And why do you have contexts in the web layer? Maybe the contexts each correspond to an external app.

Decouple. You know it makes sense.