The content of the Elixir course is covered by the Creative Commons (by-nc-sa) https://creativecommons.org/licenses/by-nc-sa/4.0/
Elixir course
This course is an asynchronous course that it is running in Taxfix to give the opportunity to developers who are not (yet) confident to Elixir to power up their skills.
This course, is also an open to contribution course! We decided to open it to
everyone to give back to the Elixir community.
We are also sharing the slides of the talk Drinking a little bit of Elixir that was delivered to initiate this Elixir course.
Episode 0
(Endtroducing...)
Hello everyone! 👋🏿
This is an hands-on Elixir Course for you folks, that are curious or already looking into learning and contributing to the Elixir ecosystem.
The scope of this course is not to teach you how to do:
iex> 1 + 1
2
iex> IO.puts "Hello Elixir world!"
"Hello Elixir world!"Ops, we already did it!
Instead, you as coders can independently check the elixir documentation and skim through the basics.
But let's step back and do an introductory TL;Do Read section!
Before talking about Elixir, we should talk about…
Erlang
Because Elixir has vintage roots.. for instance Erlang was created in 1987 and we have to know our past to know the future (or better, the present!) So..
Erlang is the “father” of Elixir
And by definition..
Erlang is a programming language used to build massively scalable soft real-time systems with requirements on high availability. Some of its uses are in telecoms, banking, e-commerce, computer telephony and instant messaging. Erlang’s runtime system has built-in support for concurrency, distribution and fault tolerance. Erlang runs on the BEAM
mmm.. so… What is the BEAM?
The BEAM is the virtual machine at the core of the Erlang Open Telecom Platform (OTP). BEAM is part of the Erlang Run-Time System (ERTS), which compiles Erlang, Elixir, Lisp Flavoured Erlang (LFE) (and more..) source code into bytecode, which is then executed on the BEAM
What about Elixir?
Elixir is a child language of Erlang that runs on the BEAM
More formally, Elixir is…
…a dynamic, functional language designed for building scalable and maintainable applications.
Elixir leverages the BEAM (Also called sometimes Erlang VM), known for running low-latency, distributed and fault-tolerant systems, while also being successfully used in web development and the embedded software domain.
Elixir is an immutable and functional language focused on concurrency
You can pass state from a function to another, however if you have the need to keep state in memory how you do it?
You can...
"Assign" the result of the function to a variable! For example let's assume that we have an append function that puts at the end of a list an element:
iex> my_list = [] # "Assign" or bind an empty list to my_list variable
[]
iex> append(my_list, 1) # appends but not "assigns" to a variable
[1]
iex> my_list
[]
iex> my_list = append(my_list, 1) # Append to my_list "1" and assigns
[1]
iex> my_list # Check if the value is there
[1]But.. there's another way!
So, let’s talk about processes…
In Elixir, all code runs inside processes. Processes are isolated from each other, run concurrent to one another and communicate via message passing.
Processes are not only the foundation for concurrency in Elixir, however they also provide the building blocks for constructing distributed and fault-tolerant systems.
Bear in mind that processes are NOT always necessary even if they are very cheap to create. You can create thousands or millions of them very quickly. As an example, a process in Elixir can be compared to the speed of checking out a new git branch.
If you are curious about when processes can be used check this out https://hexdocs.pm/elixir/master/library-guidelines.html#avoid-using-processes-for-code-organization (hint: read it ;) )
Tasks
Tasks are a way of achieving concurrency in Elixir, and follow a similar async/await pattern like in other languages. Ideally you don't want to be communicating with other processes from inside the task, as this can get messy and lead to things like deadlock. Instead the approach that is usually taken for more complex tasks, is to use a GenServer.
With this in mind, there are use cases for Tasks, so it's worth having a read about them https://hexdocs.pm/elixir/Task.html.
GenServer
Whenever you read about Elixir, you're sure to read about GenServers. GenServers are nothing more than a process that implements a basic client-server behaviour and allows you to run tasks asynchronously and store state. GenServers contain a mailbox where they store their incoming messages and process them as they come. Due to this, it's better to have multiple GenServers if possible rather than just one, as GenServers can quickly become a bottle neck for your application. It's totally acceptable (and expected) that you could have one (or more) GenServers per user.
Developers new to the language can be quick to jump to conclusions about a problem, and start using GenServers everywhere. In practice, it's better to only use GenServers when you really need them (e.g storing state and working on that state). If you don't need a GenServer, don't use one. Just because it's an Elixir program, doesn't mean it needs a GenServer.
In cases where you just want to store state, and access and update it. It's worth looking into Agents.
OTP (Open Telecom Platform)
OTP is a collection of useful middleware, libraries, and tools written in the Erlang programming language. It is an integral part of the open-source distribution of Erlang. The name OTP was originally an acronym for Open Telecom Platform, which was a name for marketing before Ericsson released Erlang/OTP as open source. However neither Erlang nor OTP are specifically designed to build telecom applications only.
Keep in mind that Erlang/OTP and Elixir they give you the lego blocks to build fault tolerant distributed systems, however creating distributed systems does not come for free with OTP. Although compared to other programming languages you get a rich ecosystem to reach this goal! (creating distributed systems)
Let’s go back to talk about the BEAM (The Erlang Virtual Machine)
On the BEAM you don’t have an event loop. The way the BEAM handles pre-emption is through reductions.
So, when you have a faulty process…
….that is consuming a lot of CPU, the rest of the processes will run without impact and smoothly.
Being the BEAM pre-emptive, it introduces stability to the system and especially predictability.
If designed well, a system thru-put is predictable even if you have a high load.
Garbage collection
Garbage Collection is different from virtual machines like the JVM. Garbage Collection is applied per process and you don’t have big Garbage Collector pauses.
Happy path programming
With OTP at your disposal, you can program in a non-defensive way. In certain cases you want that, if there is a fault in one of your processes, you want to let it crash (or restart) to clean its state. In some cases you would want a process to crash (due to an unexpected error) and restore its state. There are techniques to do this, however you would have to pay attention to avoid to restore a corrupted state.
The gist of the "let it crash" philosophy is that you can't handle all the possible ways of generating exceptions. There will be always an exception that we did not think about, so a Supervisor process will take care to restart a crashed process with a specific strategy that you can specify case by case.
Surprisingly for many programmers Erlang and Elixir have exceptions handling, however they are usually used in critical parts of the code. Especially in IO handling.
Not you will dive into episodes that will give you a taste of Elixir!
Suggestions on how to approach the course..!
As you write code in a new programming language some people are the retype code folk to train muscle memory or the copy&pasta folk and understand after what you did, etc...
Choose your style!
It's not important to be quick to finish, it's not a race! ;)
Take your time to understand what you are doing.
Check the Elixir documentation, try to go deep, but don't get lost! ;)
Links to check:
- The official Elixir tutorial: https://elixir-lang.org/getting-started/introduction.html
- Elixir School, a very good resource to quickly catch up with Elixir details https://elixirschool.com/en/
- The Elixir documentation (check this out when you have any doubt!) https://hexdocs.pm/elixir/Kernel.html
- The Elixir Forum is a super good resource to find answers or participate in (public) conversations https://elixirforum.com/
- Comparison of Erlang Runtime System and Java Virtual Machine (you might even learn about java
🙀 ) https://pdfs.semanticscholar.org/aeee/adf0c2a2dc453c0e9ef18a0239e9e796e497.pdf?_ga=2.69208912.1940899817.1594286572-947597467.1594286572
Episode 1
(Wet your appetite..!)
You will build a small textual game in Elixir, inspired from a puzzle game from Valve (https://en.wikipedia.org/wiki/Portal_(video_game))
You will start to experiment with basic Elixir syntax through the iex REPL
- Learn basic pattern matching
- How to create anonymous functions
- You will learn that in Elixir there are just modules and functions
- How use the built-in Elixir code formatter
- What is an Elixir protocol
- You will learn about the Supervisor OTP behaviour and how to start dynamically processes and Supervision trees
- You will understand how to store state in a process (By using the Elixir Agent behaviour) and transfer the state of a process to another process in two different local erlang nodes running on your laptop.
- If you would have the ip address of someone else you could create also an erlang mesh between two remote laptops!
Start from here and take your time! https://howistart.org/posts/elixir/1/
Episode 2
(Sailing the seas of functions)
In this episode after holistically 🧝🏿♀️ wetting our appetite with the portal game, we'll fallback to only plain Modules and functions.
We think that having a good command of Modules and functions will help you to reason about code (without thinking about processes from the start)
As aforementioned in the introduction, processes are super cheap to create in erlang/elixir, however they are not meant to be considered like objects in the OOP sense.
You could refresh this concept by skimming again through https://hexdocs.pm/elixir/master/library-guidelines.html#avoid-using-processes-for-code-organization
To dive in the seas of functions (and Modules!) we will use https://exercism.io/
Exercism let you learn a new language (not only Elixir!) through small challenges tackling different problems. Some challenges are easy, fun, and relaxing. Some are a tougher. But you can choose what do you want to do.
Every challenge has a set of failing tests that you have to make it successfully pass by gradually implementing the solution (or in one shot if you are very brave!)
Everyone in an exercism team can review the solution to a challenge of the insert coin team or get inspired by other insert coin team members solutions
Use this link to use to join the "insert coin" exercism team
https://teams.exercism.io/teams/NDQGCdvzZkK56nfMtNVqWdYg/join
Episode 3
Releases
Now that we know how to build a basic project, we need to know how to release it. Historically releases have been a bit of a pain in Elixir, but luckily for you, the newer versions of the language come with a built in task!
https://hexdocs.pm/mix/Mix.Tasks.Release.html
Mix release offers us a very quick way to build a release of our application. However, if you've decided to get fancy with your applications and build applications that are stateful, it's worth spending some time learning about Hot code releases in Elixir, as well as a tool called Distillery, which helps to support this.
https://hexdocs.pm/distillery/home.html
Season break!
On this special episode we are going to review what we have learned so far by building a version of the Conway's Game of Life.
The Game of Life is a zero-player game, which means that it requires no player interacting with it.
The original setup consists of an infinite, 2-dimensional grid of cells, and each cell can be alive or dead. At each step of time (tick), the following rules apply to decide what happens with each cell.
- Any live cell with fewer than two live neighbours dies, as if by underpopulation.
- Any live cell with two or three live neighbours lives on to the next generation.
- Any live cell with more than three live neighbours dies, as if by overpopulation.
- Any dead cell with exactly three live neighbours becomes a live cell, as if by reproduction.
And it looks something like the following (but this is a finite 2d grid)
And ours will look more like this (just the looks, it is not actually following any of the rules).
You may follow your heart for that or these other instructions
Bot Project
For this project we want you to build a basic bot using Hedwig (https://github.com/hedwig-im/hedwig). Hedwig provides the basic framework for setting up your application, so you can follow the instructions on the GitHub page to get started. You're totally free to decide what you'd like your Bot to do, some examples of things that have been done before can be found here https://github.com/enilsen16/awesome-hedwig.
Episode 4
Erlang Term Storage (ETS)
ETS is a storage system built into OTP that allows us to store Elixir and Erlang objects directly and offers constant time access. You can think of ETS as a caching mechanism, as the tables are killed if you take the application down. If you want something similar to ETS that offers longer term storage it's worth looking into dets or mnesia.
Ecto
Ecto is a database wrapper that is widely used in the Elixir community. For version 3 Ecto was split into 2 different libraries, :ecto and :ecto_sql. :ecto_sql is required if you intend on working with a database, otherwise you can just use :ecto to create schemas (schemas that don't back up to a database are called embedded_schemas) and take advantage of the validation opportunities that it provides.
Phoenix + Phoenix Live View
Phoenix is to Elixir what Rails is to Ruby, arguably Phoenix has been one of the greatest draws for people to the Elixir community. Phoenix offers all that you would expect from a framework such as Rails or Django, but has the advantage of being easily scalable as well as being easily able to achieve concurrency due to being built on top of Elixir.
Phoenix Live View is a relatively new addition to Phoenix, and enables users to create server rendered templates that can compete with the likes of React by being able to send partial state updates to the user. Compared to other frameworks where whenever an update happens, the entire page is sent, Phoenix Live View offers incredible performance as well as ease to setup and get started with. Take a look at the video below to get an idea of what you can achieve with Phoenix and Phoenix LiveView.
Project Suggestions + Ideas
For the final stage in the course we'd like you to create your own project that takes advantage of Phoenix + Ecto. There's no other requirements, other than to use these two libraries (if you install Phoenix, Ecto comes included
Project Ideas (they are just proposals)
- Markdown Editor - Build a site that has two panes, on one side you write markdown and on the other side it automatically renders it.
- Job Scheduler - Build an application where you can define various background tasks, and then create an interface to see the progress of your tasks, as well as being able to start, restart and pause a task. For task management you could use GenServers or take a look into a library called Quantum.
- Chat Application - Take advantage of Phoenix's great websocket support and create a chat application that allows multiple people to talk across various channels. If you want to take it further you could support message history and private messaging between users.