Unpacking Elixir: Phoenix
underjord.io
underjord.io
For example, you can use Ecto Schemas and Changesets to validate the shapes of your API requests (down to the valid values, etc) and provide good quality error messages. And changesets are easy to extend too.
Ecto is probably my favorite part of the overall Elixir ecosystem. Phoenix and LiveView have been nice to work with (But still don't quite feel right, especially event handling and bubbling up of events up components, I wish was smoother).
Overall we still really like working with this stack and the community is great too, in the general quality of libraries available.
Could you recommend how I should get started with erlang/elixir?
https://www.manning.com/books/elixir-in-action-third-edition
From what I have seen, it’s just another programming language. I’m not getting where the large speed up is…
The dev tools (build, deps, project generation) are not necessarily better than everything else but they are very good.
The paradigm of Erlang that Elixir builds on is outstanding for building reliable, efficient services. It makes trade-offs for that. Trade-offs that mostly don't have downside for your average web app.
Things that begin to get complex in many ecosystems. Low latency "live" functionality and lots of concurrency are trivial and supported from the ground up.
It is a high level dynamic language. Think abstraction-level Python or Ruby. But it has these very different trade-offs.
Docs are good. Ecosystem is not as frantic as JS.
Phoenix LiveView is particularly in service of delivering things quickly. Follows a very 80/20 approach. High abstraction level. Covers most needs. Not shy about making trade-offs to avoid complexity. Much like Erlang originally.